Please Complete This Form And Send Us Back Via Fax Or Email: Fill & Download for Free

Twilio is a company founded by developers for developers! Twilio provides a software-based platform which enables customers to easily add voice, messaging and video to their apps. Twilio is NOT selling a final product that can be consumed by an end-user (e.g., prepackaged software to solve their business problems), but rather is providing developers/customers with the prefabricated building blocks (i.e., APIs) needed to build any communication-based functionality they desire right into their application. By just virtually buying a phone number and swiping their credit card for potential per minute usage, developers can build contextually relevant communications by preventing their own customers’ users from leaving the application when they need to interact with someone (think why WeChat is popular in China). On this side of the world, popular use cases include ridesharing apps enabling anonymous communication between passengers and drivers and e-commerce companies sending automated delivery notifications or promotional messages.So how does it work?A developer signs up for Twilio, chooses a local virtual number (e.g., with 415 area-code for San Francisco, 212 for New York City) to send and receive voice, or SMS messages. The developer then maps the virtual number to a ‘request’ URL (the application’s URL, which Twilio would request from the developer’s application server when receiving a voice call on behalf of the developer/customer). The URL, or Uniform Resource Locator, created by the application developer, would describe to Twilio how to control the content of phone calls.The developer defines a set of business rules, or instructions to handling incoming and outgoing calls for each customer cohort. These instructions include: 1) Say - inform the customer that his or her order has been processed, or play a prerecorded sound file (message or music); 2) Gather - collect information from the caller; 3) Record - record the call; 4) Reject - hang up; and 5)Dial - dial this specific rep number to forward the customer to, or set up a conference call. In essence, on one end Twilio ‘ingests’ the phone call or the message flow, while on the other, it provides APIs for developers who, in turn, instructs Twilio on how to handle the incoming or outgoing phone calls. An API, or Application Programming Interface is a prefabricated block of software code that performs basic, reusable functions—e.g., displaying text on a computer screen, enabling inter-app communication—to allow developers to simply focus on building value-add, user facing applications.The call gets ‘load-balanced’ to a number of nodes/servers in Twilio’s cluster (managed by AWS). Each group of machines/servers perform specific roles—for example, some roles might be CPU-bound, a task is determined by the speed of the central processor, while others are memory-bound, determined by the amount of memory required to hold data—allowing Twilio to scale each group of machines independently of one another. The call comes in, gets load-balanced to a ‘Voice’ node cluster (i.e., a group of servers dedicated to voice function) to be answered by Asterisk, an open source PBX software platform.In general, a PBX allows telephone users to set up circuit-switched calls to other users in the same company (without toll charges) or to connect with users of the public telephone network. Effectively, PBXs shift some of the switching system out of the telephone company’s central office ‘CO’ and into the customer’s premise. The switch permits direct inward dialing ‘DID’ to a specific extension. Twilio stores the ‘DID’ numbers in its database; on the other end, the developer/customer would associate her or his assigned ‘DID’ number to a specific URL of a web application. As the call comes in, Twilio would make an HTTP POST request to that URL. In turn, Twilio would receive from the application server the XML instructions on how to handle the call.How does a phone call work? Historically, the telephone network has been a hierarchy of smaller networks performing different switching functions. In the US, switches were organized into classes from 1 to 5, Class 5 being assigned to the End Office or Local Exchange ‘LEX’. Classes 1 to 4 were Toll Centers or Transit Nodes ‘TN’. For a particular country, this network hierarchy would also be connected to international networks through a higher switch called International Gateway. Even if the modern structure has flattened, removing the Class 1 to 3 switches from the hierarchy, Class 4 and Class 5 designations remain in use today.The core of a telephone network is made of switches—TN and LEX. The primary function of these switches is to establish connections between telephones and the transmission equipment used to carry the voice call. When a called is placed, the switch detects that the receiver has been lifted, provides a dial tone and collects the called number. It then translates the dialled digits into a destination path by finding and reserving an idle circuit—if not it returns a busy-line tone. When the connection is established it sends a ringback tone to the caller until the destination picks up the phone. Then the connection is established by linking all the connections previously reserved in the transmission network and in the end terminates the call by releasing network resources reserved for the call.Typically, residential customers are connected to switches through an Access Node ‘AN’ or a Digital Loop Carrier ‘DLC’. This equipment converts the incoming voice into the appropriate digital format and also performs some basic call-processing functions. DLCs are access solutions that enhance the reach of the switch for remote customers. They do so by multiplexing many customer lines into just a few trunk lines that return to the switch.How does an internet connection work? The main historical way to access the Internet was a dial-up connection or narrowband access. Narrowband access uses the phone line to the Local Exchange, which is then connected to the data network through a Narrowband Access Switch ‘NAS’. The data network is typically a set of IP and/or ATM switches. The most widespread broadband internet access today is DSL. It re-uses the telephone copper wire up to Digital Subscriber Line access Multiplexer ‘DSLAM’ that directs the voice traffic onto the PSTN and the data traffic to the data network through a Broadband Access Router ‘BAS’. The PSTN, or the Public Service Telephone Network is a circuit-switched network that sets up dedicated voice circuits across a network of copper and fiber optic cabling.How do converged networks handle calls? There are many different ways to converge networks. In general, Network convergence requires the addition of two types of network elements: Media Gateways (MG) and Softswitches. Media gateways act as the physical bridge between the two networks. Their main purpose is to convert the TDM voice packet into data packets—generally IP or ATM. The softswitch acts as a media gateway controller by extending the voice network signalling ‘SS7’ into the data domain.When a phone call is placed on such a converged network, both the voice connection and the signalling are routed to the data network through a media gateway. The softswitch is a central database ensuring that Media Gateways properly communicate between each other.How does Twilio Fit in? To properly understand the inner workings of the Twilio platform, it is important to understand how basic web browsing work.What happens when you type a URL into a web browser. When you type aws.amazon.com for example, into a web browser, a behind-the-scenes process occurs translating the domain name (aws.amazon.com) into the IP address (54.239.31.69). The IP address is sent to the web browser so the user can be connected to the website. This process, called resolution, relies on a global network of name servers. The browser is a program that requests information from a web application server, which in turn finds the requested information and sends it back to the user’s client (desktop, mobile).A web server is essentially a software program that sits in the middle tier of an ‘N-Tier’ computing environment between client-side environments (web browser), and server applications. Application servers allows developers to focus on creating specific business logic objects without having to code at the system level. The primary role of the application server is to access enterprise servers (e.g., AWS’ servers) for business logic and data, serving application objects to clients in the form of static or dynamic HTML pages, JavaScript, etc.Clients and servers communicate using Hypertext Transfer Protocol Language (HTTP). IP addresses are what machines on the Internet use to identify one another. It is a 32-bit address defined by Internet Protocol, or IP that uniquely identifies each computer on the Internet. Every computer, or machine on the Internet has a unique IP address. If a machine does not have an IP address, it is not really on the Internet. When searching for frequently visited Websites (e.g., aws.amazon, Facebook, Google, etc.), the recursive name server already has the information cached in its memory and passes the IP address on to the user’s computer browser without having to take any further steps.The ‘conversation’ taking place between the browsers and the local recursive name server is analogous to the following dialogue:User’s computer: “What is the IP address of the URL Amazon Web Services (AWS)recursive name server: “I already know that. The IP address is 54.239.31.69.” The local recursive name server then delivers the address to the end user’s browser.Under ‘the cover’, a ‘load balancer’ differentiates between a request to a main page and another to its extension (e.g., www.amazon.com and today’s deal extension www.amazon.com/gp/goldbox ),and sends the request for the web page of today’s deals to a group of servers that is optimized for that task. A load balancer goes one step further by allowing traffic routing decisions to be based on the so-called HTTP header (HTTP, or Hypertext Transfer Protocol is an application-level protocol that is the foundation for the World Wide Web). A load balancer’s controller not only examines the URL, but also information such as cookies, client source address, etcTwilio Client SDK enables VoIP communications. Twilio Client enables developers to integrate Voice Over IP, or VOIP into their applications.To illustrate, let’s examine a scenario whereby hypothetical InsuranceCo were to roll out Salesforce CRM Call Center mobile application through which is provisioned several customer service numbers (local numbers, international, 1-800 etc.) provided by Twilio. InsuranceCo buys a virtual 1-800 number from Twilio. InsuranceCo’s customer calls the 1-800 number. Salesforce has set up workflows, on behalf of InsuranceCo, for how to deal with inbound calls. Twilio Client SDK enables VoIP communications within the Salesforce application. The customer’s carrier (e.g, Verizon) routes the call to a Competitive Local Exchange Carrier ‘CLEC’ with whom Twilio has contracted (e.g., Level 3) . The call is routed through AWS to Twilio, triggering an HTTP request to Twilio’s call API to initiate the call.On the other end, Twilio 1) looks up the voice URI (Uniform Resource Identifiers) of the Salesforce Service Cloud application, and 2) provides Salesforce’s web server +information about the caller (e.g., originating and dialed phone numbers, date and time of the call, geographic location of the caller, etc). Twilio would also request instructions from the Salesforce application server on how to handle the call. The process is quite similar to how a browser makes a request to a web server discussed above.Once the Salesforce application receives the caller’s information (i.e., the call is logged into Salesforce). Processing function follows, involving a caller/customer database lookup to map customer’s caller ID to her or his historically stored information (e.g., caller's support ticket history). Caller’s information are then injected into the Salesforce application workflow. In this example, InsuranceCo.’s marketing database (maintained by Salesforce) has been capturing data about the customer’s behavior from various sales and communications channels.Twilio’s TaskRouter. Salesforce’s application server then sends to Twilio’s server (which is hosted on AWS) XML instructions on how to handle that specific phone call (e.g., since the analytical engine determined the caller is a high value distributor/customer, please forward immediately the call to highly skilled call-center rep A). An HTTP request is sent to a virtual call router’s API (i.e., Twilio’s TaskRouter's API). The call router, or TaskRouter authenticates, or looks up the account holder’s unique ID, matches the caller to the intended rep, then sends an HTTP request to Twilio’s server to initiate the phone call between the caller and the rep.The existing data on the caller/customer is enhanced by the overlay of the real-time contextual communication data elements; i.e., putting the customer call in context of the Salesforce’ customer service application, providing the call center agent real-time intelligence, needed for improved customer profiling. The agent can instantly access and view prior communications, use real-time analytics identify the personalized best next-action for the customer (e.g., promote a special offer, provide a service alert), and through, for example, Pitney Bowes’s content-creation platform, automatically integrates incoming data, compose new offer and send it to the customer through the relevant channel.What is Voice over IP ‘VoIP’? Voice over IP (VoIP) is a technology that involves sending telephone calls over data networks, such as the Internet, rather than the traditional Public Switched Telephone Network ‘PSTN’. Traditional phone calls across the PSTN use a dedicated circuit that transfers calls as uncut streams, allowing no other information on the circuit regardless of available bandwidth. Conversely, IP networks transfer data more efficiently in packets that get reassembled on the receiving end rather than using a dedicated circuit. For a VoIP call, audio is first converted from an analog signal to a digital signal through a codec, separated into discrete packets, sent across an IP-based network, put back into order at the termination point, and then converted back to an analog signal to create audio that the end user can recognize. There are three major VoIP protocol standards—namely, H.323, SIP, and MGCP.The Session Initiated Protocol, or SIP In March 1999, the Internet Engineering Task Force ‘IETF’ introduced Session Initiation Protocol ‘SIP’, which was designed to support quicker call set-up times and enhanced Web capabilities. SIP does not require gateways to maintain all call information during the life of the call and, as a result, it is less demanding of servers and is more scalable.SIP focuses on session initiation, modification and termination, and leaves the session and connection details to be negotiated by the end points. SIP uses a simple text command structure with HTTP syntax and URL addressing. Thus, SIP is well suited for any Internet- and Web-based applications.SIP allows direct communication between clients via a peer-to-peer ‘P2P’ connection using only IP addresses, and in most cases a SIP proxy will query a Domain Name Server ‘DNS’ to resolve a domain name into an IP address of record. An important characteristic of the SIP protocol is that it allows endpoints to establish a peer-to-peer connection and communicate directly. This is in contrast with other protocols that require resources on the network to control communications between end points.How does SIP “Work”? To initiate a phone call, the SIP phone issues an INVITE containing the caller’s IP address and the type of media requested (e.g., voice). From there, the SIP servers (Proxy and Redirect server) send back the SIP-URI where the called party can be reached. In this respect, the SIP servers somewhat resembles the DNS, or Domain Name Server servers on the Internet.What is the Domain Name System ‘DNS’? The Domain Name System (DNS), a global, distributed database infrastructure, is part of the fabric that holds together the Internet—performing the simple, straightforward function of mapping URLs (Uniform Resource Locator) to IP (Internet Protocol) addresses. Every Web server on the Internet has one or more unique IP addresses, represented as four sets of numbers, called octets, separated by periods (e.g.,140.211.169.9). The DNS enables people to use domain names (e.g., www.opendaylight.org ), which are much simpler to remember, to identify Web servers as opposed to IP addresses (e.g., 140.211.169.9). Each time a user enters a domain name into a computer’s Web browser, the DNS translates the user-friendly URL into the IP address needed to locate the appropriate Web server.SIP uses the above request and response method to establish communication among the various network components and ultimately enable a multimedia conference between two users. Users are identified by globally reachable unique addresses called URIs. URIs use the same format as email addresses to describe IP service points (e.g., [email protected]). Users register their assigned URIs with the registrar server, which provides this information to the location server upon request. Users can have multiple URIs with different service providers that point to the same device, but they can also be reached with traditional telephone numbers. Calls using these traditional numbering schemes are translated into SIP URIs using the ENUM method.E.164 Number Mapping ‘ENUM’: Telephone Numbers on the Internet ENUM, or E.164 Number to URI Mapping translates telephone numbers into Internet addresses. ENUM is a protocol that merges the Public Switched Telephone Network ‘PSTN’ and the Internet—mapping complete international telephone numbers to URIs. Since the SIP protocol is IP based, it provides users (and applications) globally reachable addresses called URIs (Uniform Resource Identifiers). URIs use the same format as email addresses to describe IP service points (e.g., tel: [email protected], mms: [email protected], etc.) and can be reached with traditional telephone numbers. Calls using these traditional numbering schemes can be translated into SIP URIs using the ENUM methodology.To put ENUM into context with the aforementioned technologies, SIP performs the initiation of interactive communications sessions between users, as well as termination of those communications and modifications to those sessions. SIP is one protocol that may be used by ENUM to send out initiation attempts to multiple locations in order to find the user who is receiving a call. By placing telephone numbers into the DNS, ENUM can facilitate a range of applications including addressing for fax machines, email, instant messaging, etc.What Value Does ENUM Add? ENUM enables users to access Internet services from Internet enabled phones, ordinary phones that are connected to Internet gateways or proxy servers and/or other Internet devices that may have inputs limited to a numeric keypad.ENUM also provides users with greater control over communications, including allowing users to indicate their preferences for receiving communications. For example, a user can indicate a preference for voice mail messages over live calls during certain times, or may specify a call forwarding location.ENUM allows an end user to reach an IP device by dialing a telephone number rather than entering a URI. A traditional number is entered into the calling device, and the number is then transformed into a fully qualified address by an application or a device that supports ENUM.A developer’s customer dials a Twilio provisioned virtual phone number. The SIP proxy queries the ENUM DNS server to resolve the fully qualified domain name into a URI. The SIP proxy will then query a DNS server to determine the IP address to send the invite. To illustrate, let’s walk through a scenario whereby End User A dials a Twilio’s business customer, or User B:User A dials User B’s phone: +1-646-470-8021.Internet Gateway converts number to a domain name and queries VoIP local recursive name server: 1.2.0.8.0.7.4.6.4.6.1.e164.arpa. By using ENUM, e.164 numbers can be used in DNS by transforming the phone number into a hostname. This is simply done by reversing the numbers, separating each digit by a dot and then adding the e164.arpa suffix.13 For example, the number +1-646-470-8021 would be transformed to the fully qualified domain name 1.2.0.8.0.7.4.6.4.6.1.e164.arpa.Local recursive name server: “I don’t know that address, but I’ll check with a name server that does. Hold on for a millisecond.”e164.arpa Tier 0 server: “Here are the addresses for the authoritative name servers for the CC1 1.e164.arpa domain.”domain.com name server: “The URI for 1.2.0.8.0.7.4.6.4.6.1.e164.arpa is [email protected].”User A’s telephone contacts User B’s (Twilio’s customer’s) telephone at returned IP Address. The local name server launches a query to the DNS, which responds with the IP address (e.g.,108.231.245.239) of the local proxy server associated with [email protected]. The SIP proxy server in User A’s network contacts the SIP proxy server in User B’s customer network, and the proxy server in User B’s network then contacts User B’s destination SIP IP phone.When the called agent receives the INVITE request, it determines if it can accept the call. If it can, it will ring the phone and then send back a response to let the calling agent know that it is ringing. When the phone is answered, User B’s, or Twilio’s business customer’s phone sends an OK response back to the calling agent along with its media capabilities. Both agents agree on a media channel, and User A’s phone sends an ACK to User B’s phone.After responses and acknowledgments are exchanged, an RTP ‘Real-time Transport Protocol’ session is established between SIP IP phones of Users A and B, or the end-user and Twilio’s business customer.Traditional Voice and the Public Service Telephone Network ‘PSTN’. The voice telephone systems are referred to as the PSTN ‘Public-Switched Telephone Network or POTS ‘Plain Old Telephone System’. PSTN is a circuit-switched network that sets up dedicated voice circuits across a network of copper and fiber optic cabling. The structure of the traditional telecommunications industry varies by country and depends on the nature of the regulatory environment.In the United States, the industry has been pushed into a competitive model consisting of a variety of participants, primarily oriented around local exchange carriers ‘LECs’ that provide last-mile connection for consumers and businesses within specific geographies ‘LATAs’, and Interexchange Carriers ‘IXCs’ that provide long-distance services. A more graduated categorization includes ILECs ‘Incumbent Local Exchange Carriers such as SBC/Verizon), CLECs ‘Competitive Local Exchange Carriers’ (e.g., Level 3), IXCs (e.g., MCI/Verizon and AT&T), and ISPs ‘Internet Service Providers such as Earthlink and Prodigy/AT&T.Twilio interconnects to the so-called Tier 1 carriers - carriers that own or control sufficient portions of their underlying network infrastructure;e.g., Verizon, Level 3 - to provide the PSTN integration. It uses SIP (described above) origination and termination to ‘talk’ to originate and terminate calls on the PSTN.Traditional PBX. As discussed above, a PBX is a telephone switch located on the premises of a company. Traditional PBXs were typically hardware-based solutions that ‘sat’ inside customer premises ‘CPE’, providing businesses with the benefits of direct dialing, call forwarding, and a variety of enhanced services. Put another way, PBXs were originally designed for businesses as a cost-effective alternative to the provisioning of individual lines to each end-user from the phone company’s central office. The PBX is like a mini-CO, owned and operated by the corporation itself. In this respect, traditional PBXs reduced both line provisioning costs for the corporation and telecom services expenses associated with intra-office calls.Mostly Proprietary. Nonetheless, because PBXs are highly proprietary systems, enterprises have had to rely heavily on the PBX vendor to deploy or integrate any new applications. In a traditional PBX, there is a proprietary operating system running on a computer processor in a proprietary chassis. The applications are also proprietary, running on the same or separate processors. The interfaces—trunk cards and line cards— are also proprietary. In short, the traditional PBX is like a black box, with the vendor controlling virtually everything—the addition or adjustment of applications generally needs to be made by the PBX vendor. The proprietary nature of the technology is often predicated by its complexity. In fact, a high-end legacy PBX usually incorporates about 5 million lines of code.Limited Scalability. One of the most significant limitation of legacy PBX systems may be scalability. PBXs are typically designed for a specific number of users, and once the enterprise expands beyond that specific capacity, a new and bigger PBX needs to be installed. Sometimes, small businesses have to purchase a higher-end PBX than they need in case of possible future expansion, resulting in a particularly inefficient use of capital. It is also problematic to connect PBXs across multiple sites, and the signaling between PBXs is proprietary. Another key problem is that handsets which customers may have purchased for the lower-end PBXs often do not work with higher-end PBXs. As a result, customer upgrades to a higher-end PBX system often necessitate the additional cost of purchasing new handsets.Asterisk and IP-PBXs. Enterprise networking focuses on three primary goals: 1) Scalability, 2) Controlling the cost of communication through the most efficient use of technology and carrier services, and 3) Improving the productivity and performance of workers by distributing information to support their activities. With the advent of the Internet, IP PBX systems were introduced and allowed for phone calls to be placed over IP-based, rather than over TDM-based networks. In such an IP environment, distributed communications servers ‘IP-PBXs’ provide scalability and redundancy by sharing and quickly reconfiguring resources in the event of individual server failure. This redundancy and distributed processing is only feasible because the architecture separates the voice traffic from the PBX, leaving only call signaling and processing responsibilities to the PBX; hence, enabling independent scalability.Asterisk, on the other hand, is commonly used open source PBX software platform, developed in 1999 by a company called Linux Support Systems that later changed its name to Digium. The development of Asterisk is predicated on the idea that modularity, or separating a PBX system into interconnecting components—akin to a boxful of LEGO bricks—would enhance scalability. An Asterisk based IP-PBX is essentially a x86 communications server, running on Linux.Hosted PBX With Hosted PBX, PBX functionality is delivered as a service over the carrier’s network. Enterprise customers typically pay for the service under a leasing arrangement. Rather than having a PBX system located on the enterprise’s facilities, those functions are located in the carrier’s network and delivered over IP-based trunks to the enterprise.SIP Trunking involves the direct IP connection of a SIP-enabled IP-PBX and SIP-compliant VoIP service provider. It is an IP-based service provided by telecom operators (and Twilio) to connect an enterprise’s PBX with the service provider’s network. Put another way, SIP Trunking is an IP connectivity consisting of a single pipe which connects a service provider’s network to an enterprise IP PBX. As discussed earlier, SIP, or Session Initiation Protocol, is the protocol used to set up IP-based sessions between network endpoints such as end-user devices or servers. SIP trunks allow operators to provision VoIP voice sessions. Enterprises benefit from SIP trunking as it consolidates their voice and data networks, replacing premise-based connectivity, and thus reducing overall costs. In the past, enterprises had to connect bundles of physical wires—PSTN lines—to a business—PSTN endpoint. A SIP trunk eliminates the PSTN lines—reducing the number of SIP connections per port—and other associated equipment such as PSTN gateways.Thus, SIP Trunking offers a number of advantages over traditional TDM-based connectivity. First, SIP Trunking allows the enterprise to reduce its telecommunications costs. While many enterprises already save on the cost of voice calls between their sites by implementing IP-based PBX systems and using intra-corporation VoIP calling. Using SIP Trunking, enterprises can further expand their ROI by extending VoIP outside of the corporate LAN.The savings comes from: 1) Getting rid of traditional analog/POTS, ISDN BRI ‘Basic Rate Interface’, ISDN PRI ‘Primary Rate Interface’, or T1/T3 subscriptions; 2) More optimal use of SIP trunk bandwidth as both voice and data services can be delivered over the same connection; 3) Greater flexibility in purchasing voice capacity as enterprises don’t have to purchase lines in groups of 24 T1 or 30 E1 lines; 4) Flexible routing of calls to preferred carriers —long distance calls can be made for the cost of local calls; and 5) Lower operating costs of IP-PBX systems vis-à-vis traditional TDM-based PBXs.For most enterprises, the desire to save money is the primary force driving adoption of SIP Trunking. Least cost routing is an interesting example. Enterprises can utilize SIP trunks from multiple service providers and proactively route specific calls to certain carriers based on country codes—operators often charge different international rates based on availability, time zone differences, and geography.Moving beyond these types of cost-reduction initiatives, SIP Trunking enables a host of additional capabilities that enterprises can benefit from. As discussed earlier, the SIP protocol itself was designed to initiate all types of real-time communications over IP networks — not just voice. Today, enterprises are taking advantage of not only the significant cost savings afforded from SIP Trunking, but also the ability to improve enterprise productivity through the deployment of Unified Communications applications.Where Twilio fits in: Elastic SIP Trunking.Companies such as Acme Packet (acquired by Oracle) use their Session Border Controllers SBCs as SIP firewalls. A session border controller is a piece of network equipment or a collection of functions that control real-time session traffic at the signal, call control, and packet layers as they cross a packet-to-packet network border between networks or between network segments. SBCs are typically located at the perimeter of disparate IP networks, such as between headquarter and branch offices, and/or between call centers and enterprise data centers. They provide network operators with ‘policy-based control’ over VoIP sessions, furnish basic protocol inter-working and defend the carrier backbone against a variety of attacks.SIP trunks are VoIP trunks from service providers that use SIP for call control and routing,enabling enterprises to create a single, pure, IP connection to a carrier’s core network. This can be viewed as an Enterprise IP-PBX that peers with the service provider core SIP proxy. “Twilio-SIP is for use in TwiMl type applications to terminate or originate a call from a known SIP endpoint or address. Elastic SIP Trunking is utilized when you have an existing application or appliance that needs to have origination and termination capabilities (think a PBX like Asterisk/Freeswitch) and Twilio will be that provider”.Programming Messaging.Twilio’s Messaging API enables developers to embed text-based communications in their applications. Using the same virtual number, our hypothetical InsuranceCo (discussed above in the voice section) can use the same both make and receive voice calls, and send and receive SMS.SMS would follow a similar pattern of voice--the flow from caller/sender to receiver--yet with the addition of an SMS aggregator. Say for example an Uber rider contacts a driver. The Uber application would use Twilio’s API to generate and initiate the SMS to the driver’s number. The SMS is routed through AWS to a CELC (e.g., Level 3). The CELC would then route the message to an SMS aggregator with whom Twilio has contracted (e.g., Syniverse). The SMS aggregator routes the SMS to the driver’s carrier (e.g., AT&T). The driver’s carrier then routes the SMS to the driver.The SMS Aggregator. An SMS aggregator such as Syniverse (acquired by The Carlyle Group)is essentially a ‘clearinghouse’ provider that facilitates wireless roaming between different carriers’ networks. As a third party intermediary, it plays a necessary role in a complex telecommunications environment characterized by different network architectures, signaling standards, billing record formats, and network protocols. The aggregator serves as the middle hub connected to all carrier partners, allowing each to roam on other’s network (assuming a roaming agreement is in place).Continuing on with the Uber example, when the SMS is sent by the rider, a detail record is created. This detail record contains basic information about the SMS (e.g., who is the sender, where they sent, the length of the message, the carrier that authorized the message). This record is then stored in one of the several formats. For GSM, the format is known as TAP, or Transferred Account Procedure (TAP) file, while for CDMA the format is known as CIBER.The data record must now be communicated to the right partners—this is where an SMS aggregator comes in. Syniverse, for example, receives the information in its data center, aggregates the data, and distributes the information to the right carriers. The company then calculates the net obligation of each carrier based on the information detailed by the data records.The Short Message Service Centre (SMSC). SMS includes a number of distinct features, which I have highlighted below. These are made possible as messages are sent via an SMSC. The SMSC mainly acts in a similar way to a router of messages. However, it also acts as an important interface with other parts of the network and other systems on that network.In general, one of the main features of an SMSC is the ability to store and forward messages. If the receiving device is switched off, the central system stores the message until it receives a signal that the device is now switched on, when it will then deliver the message. In addition, the functions of an SMSC can include providing the senderInterface with Other Network Elements. In addition to the store and forward features, the SMSC can provide an important interface to an operator’s other applications and act as the router between these. For example, the SMSC may interact with the pre-paid billing system, location servers, user profiles and platforms for other SMS based applications.

Have you ever seen an employee get fired on the spot because of you?I wasn't actually in the building when the person was fired, but was shoved out of his way by the box he was carrying with all of his personal possessions in it, as we accidentally met in the building foyer after he was fired.The eventual firing of this person had began approximately one year prior to his actual termination.Maybe I still feel guilty and that's why this answer is so long and convoluted, either that, or I cant sleep tonight, so you've been forewarned.The project site was a Nuclear Power plant where I was the Lead Outside Plant Cabling Engineer and Supervisor for all Fiber Optic cabling installation, splicing, terminations, initial compliance testing and certification.The way upper management had set things up, our Telecom department had a pseudo set of checks and balances where final testing of the fiber optic cabling was performed by the group that installed and turned-up the electronic equipment that connected to the fiber optic cabling. My responsibility was supposed to stop at the fiber patch panels where the fiber terminated.The Telecom department consisted of multiple contractors employed by two main contracting firms, which made for a lot of “office politics" with a lot of finger pointing. Especially since the afore mentioned group that installed and turned-up the electronic equipment had unsuccessfully bid for the cabling work that my company was responsible for, and they were prepared to use any means necessary to win the cabling portion of the contract at the next bidding opportunity.From the start, the other contractor had questioned my fiber optic design as being overkill, with the future advent of wave division multiplexing, which had caused me to spend a week researching the technology and writing a report that basically stated that yes, multiplexing was a possible alternative in 10 to 15 years when the technology might be ready for actual field installation and the pricing might be financially equitable.(WDM technology was still pretty much in the theoretical stage at this point, with some lab experiments showing that it was possible, but it wasn't even developed enough to call it “bleeding edge” technology yet)The opposing contractor had a person that was my counterpart for the electronic equipment installations that I'll refer to as “Bernie” from here on out.Bernie and I both had recently been sent to a Siecor Fiber Optic school, at different times, that certified us as qualified Siecor Fiber Optic Engineers and Installers.Apparently Bernie had not paid as close attention as I had during our classes though as he continually tried to blame system failures on the fiber optic cabling that I engineered and supervised the installation of, even though Bernie had also tested the fiber optic cabling with an OTDR (Optical Time Domain Reflectometer) and with a calibrated Power Meter and Light Source and certified in writing that my fiber optic cabling met all Siecor certification criteria. This was all performed after my crew had performed the same exact testing as a redundancy. (Nuke plants are all about redundancy by the way)Let me repeat that for you, because I know this is getting convoluted.Bernie was blaming system failures on a fiber optic cabling system that he had personally tested and certified in writing with his signature that it met all of the manufacturer’s certification criteria.The 1st instance of Bernie's incompetence was when his team was trying to turn-up and test an IBM system referred to as “IMAGES" that the nuke plant was going to use for converting millions of microfiche film images to a digital format for easier storage. I may have the the equipment designations wrong here because it has been almost 30 years, but I think they were IBM 3270'S.As a “Beta" systems test we were going to bring the two closest engineering buildings on line, each of which were well short of being 300′ feet from our Data Center.My crew and myself had worked massive amounts of overtime to meet the preposterous IBM schedule set by them, because IBM had also bid on the fiber optic cabling, but their cabling design had also been rejected in favor of my own design, so they weren't exactly “my friends" on this project either. Personally I had over a hundred hours that week.My crew had finally completed our initial testing of the cabling 48 hours before the Beta test deadline, when I informed Bernie that the fiber was ready for him and his crew to perform the final testing and connect their equipment to it. That was on a Monday morning between 7:00 and 8:00 AM.By noon of that same day, Bernie confirmed that his crew had tested the fiber cables and were ready to begin turning up their electronics. I didn't hear anything else from Bernie the rest of that day.But the following morning, Tuesday, my pager goes off as I'm arriving on site and it's the customers Telecom Dept Manager, in other words my boss's boss and it has “911” added on behind it.Instead of calling the guy, I go straight to his office as soon as I make it through security and get to our building. The manager informs me that there is apparently a problem with the fiber cabling that I had reported as completely tested and certified and that Bernie and his crew had worked literally all night trying to get the IBM system on line. Yeah I'm feeling some heat at this point, we have until midnight to make our 1st milestone on a $20 million dollar project and all fingers are pointing at me if we miss it at this point, or it at least damned sure felt like it to me.As soon as I get out of the big guy's office, I page Bernie to see what the story is and he tells me that the 3270 flashes a green light showing signal receipt, but then it shuts off within seconds and they have to “reboot" the equipment each time afterwards with the same results each time.Keep something in mind here, neither myself or anyone on my crew had any training or experience on the electronics for this project, but now I'm in a position of trying to trouble shoot something that I had never laid eyes on before having seen one for the very 1st time, the day before when I finished the fiber testing and Bernie's crew were placing them in the equipment racks.Thankfully, during my nearly catastrophic learning curve with the Sumitomo Fiber Optic Fusion Splicer that I had convinced the customer they should purchase for about $60,000 I had made a friend named Leon (that's his real name by the way, wish I could remember his last name, but it's been too long) Leon was a field engineer for Sumitomo Electric's offices in Research Triangle, North Carolina.Leon quickly walked me through all the correct fiber testing processes just to make sure that I had done everything per industry standard and had me fax him the OTDR shots for those specific fiber strands being used along with our Power and Light meter readings.By now it's almost noon and the dept manager has been in and out of my section manager's office at least once an hour since the workday began.Within 30 minutes of Leon receiving my testing documentation, he calls me back and ask me what the “loss budget" was for the IBM equipment is. Me being me and still a rookie, I flat told him that I had no clue, but I would find out and call him back.My next call is to Bernie, who I ask for a spare 3270 manual so I can fax a copy of the equipment specs to Leon. Bernie suddenly seems to have a fishbone stuck in his craw and tells me he cant let me have one of the manuals for some bullshit reason, but he will find the “loss budget" criteria and send it to me, strange, but I don’t care as long as I get the information. At this point, I'll take what I can get, anyway I can get it.Bernie sends me an email that only has the lower equipment limits, stating the 3270's have a minimum of 20db of signal power received (I'm totally winging these numbers, because like I said, it's been almost 30 years ago).I call Leon back and tell him that the minimum is 20db and per the test documentation I had sent him we were showing over 40db received at the equipment. (Our total loss from end to end was less than 2 db, that included two fusion splices and two mated pairs of connectors.)Leon instantly tells me that Bernie has only given me part of the equipment specs and that we needed the maximum transmission limit not just the minimum.Once again I call Bernie and the fishbone in his craw has apparently gotten worse, because he mumbles that he will have to get back with me in a little bit.Just as I hang up the phone with Bernie, I hear a bellow from my manager's office requesting my immediate presence in his office.My manager was a retired Navy Reserve Senior Chief by the way, and apparently the navy teaches an advanced course in bellowing.My manager wants a quick update on the situation, now keep in mind, my manager had been getting some serious heat all morning over this issue and he has had my back the whole time.I give him a quick rundown of what the current status was and that I'm waiting on Bernie to give me the equipment specs, but that Bernie seems to be having a problem providing the information.My manager, tells me to turn around and look on the top shelf of his book case and what do I see there, why nothing but IBM equipment manuals for every piece of equipment being installed as a part of the project.Within 5 minutes I was back on the phone with Leon and told him that the upper limit was 40db of power received. Leon and I both realized that our transmission level was over the maximum specified, but whereas I had no real clue what the consequences or solutions were Leon did.Leon told me that our fiber installation was actually too good since the equipment was designed to expect a certain amount of loss over the fiber, splices and connectors, but we were so far under those limits, that the receiver was being saturated with too much light and was shutting down to prevent possibly damaging itself.OK, I understand the signals too hot, but how do we fix the problem?Leon asked me if we had any “index matching gel” (think of a lighter KY Jelly, but it doesnt dry out as quickly) on site.During my Siecor training, they had touched on index matching gel and that it was typically used to cut down on “back reflections" between a mated pair of connectors. (Kind of similar to having two mirrors facing each other, the gel helps attenuate the back reflection between them)I told Leon that we didn’t have any on site and the only possibility of anyone having any was one of our suppliers in the DFW area, which would be 60 to a hundred miles away, if they even stocked it, otherwise it would take at least 24 hours for me to have it overnighted.Leon's next comment baffled me when he said “nose grease". Of course I had to ask him to repeat himself and he said “take the fiber connector that plugs into the fiber cross connect panel and rub it on your nose to get some of that oil on it for a temporary fix until you can get some fiber attenuators (basically a small flat washer that puts an air gap between the connector surfaces and weakens the signal) on site.Leon then went on to tell me that we might need to coil the fiber jumpers around pencils between the patch panel to add “microbends” to the jumpers and equipment to add more attenuation if the “nose grease" wasn’t enough.Leon told me to give both those methods a try if needed and call him back to keep him updated.On the way to the Data center, to find Bernie, who still hadn’t called me back or emailed me the requested information, my manager and the Dept manager fell in step with me for a running update.All I told them was that I thought we might have a temporary fix until we could get some fiber attenuators on site, because we were saturating the optical receivers.Bernie wasn’t in the Data center, so I unplugged the fiber jumpers and began rubbing them on my nose as both managers looked at me like, “Oh hell, he done lost his ever loving mind". I gave them both a shrug and and a wink, then invited them to follow me over to the 1st engineering building to see if my crackpot solution would work.Bernie and one of his helpers were sitting in the buildings equipment room drinking coffee and perusing the latest copy of Playboy when I swung the door open and entered followed by the two managers who had been joined by Bernie's manager along the way.As Bernie watched in disbelief, I unplugged the fiber jumpers from the patch panel and rubbed them on my nose, plugged the connectors back in and asked Bernie to try the equipment again.We had to wait a few minutes for Bernie's helper to run back to the Data center to reboot that piece of equipment, once he radioed Bernie that his end was transmitting, Bernie restarted the 3270 on our end and instantly had all green indicator lights come on and stay on this time.I handed Bernie a post it note with the part numbers for the attenuators that were needed for the permanent solution without saying a word, turned and walked out of the equipment room without another word and returned to my cubicle.My manager stopped by on the way to his office and told me that there was an “impromptu” project meeting starting in the conference room in 15 minutes that required my attendance.This has gotten way longer than I intended, lets just say that the only question I was asked in the meeting was if I had or had not ever received a set of the IBM equipment manuals at the start of the project, to which I replied that I had not received any equipment manuals and I was told that my presence was no longer required in the conference room, nor was my manager's presence required.As my manager and I were getting coffee in the breakroom, we could hear the Dept manager chewing some ass in the conference room, you see the dept manager was in the same Navy Reserve unit as my manager and he was the next in line for Senior Chief.[EDITED portion follows, apparently I inadvertently hit submit last night before I had finished typing the full story of Bernie's demise. What follows are two more instances of Bernie's incompetence and chicanery if you haven't been put to sleep yet, read on.)Bernie's next faux pas involved the final testing of a separate fiber optic cabling system that was a critical system for the plant's Emergency Response Facility (ERF).The ERF was located in the Main Nuclear Operations Support Facility and training building approximately one mile from the plant's containment structure just inside the property's main entrance.The ERF was a back up facility that was connected via fiber optic and copper cabling to all of the plant operations systems that were typically monitored and controlled from the plant's Main Control Room located inside the containment structure. In case of a plant emergency, ERF teams could monitor plant conditions without interfering with the actual on duty reactor operators or shift supervisor, allowing them to concentrate on the emergency.The ERF team could provide needed consultations, logistics support, implement and coordinate emergency response plans with local, state and federal agencies.So the fiber and copper cabling systems connecting the ERF and Main Control Room were kind of a big deal, big enough that if the system was down for more than 72 hours, the plant owner could be fined $10,000 per day by the NRC.My fiber crew had placed and terminated a new fiber cable to replace the existing fiber cables and I had personally tested the cabling per standard procedure, so I was very confident that it was well within the system specifications.Bernie's test and turn-up crew were notified that the new cabling system was completed and ready for their final testing before system turn up on the new cabling.Within an hour of Bernie's crew beginning their testing, Bernie contacted me to tell me that their were major issues on the new fiber cabling, that due to some high reflective signatures on their OTDR shots, there appeared to be breaks in the fibers.I was already tied up on a different issue and after quizzing Bernie about the characteristics of the anomalies, told Bernie that per his own comments, that there was no way in hell the fibers were broken, or they would be showing at least some amount of signal loss shown instead of just a spike in the reflections shown on their OTDR shots.Bernie comes up with a wildly speculative theory that the reflections were being caused by the “fiber fusion heat shrink protectors” that I had specified to prevent our fusion splices from possibly being broken by plant vibrations. When I asked him where he had gotten that idea from, his reply was that he had “read an article somewhere that discussed possible back-reflection issues", but he couldn’t recall where he had read the article at.I was already at my limits and told Bernie that he was “full of shit” and that me and my fiber crew were already tied up on another high priority emergency in a different part of the plant, and that the problem was with their testing, not the actual cabling and terminated the phone call.Within 5 minutes of Bernie's and my phone call, my pager goes off again with the dept manager's number followed by 911 again.Bernie has had his section manager escalate the latest issue to the dept manager saying that the new fiber cable is failing and cant be used. The dept manager's exact words to me were “you've got a major issue that you need to address right the fuck now, drop whatever else you are working on and get your fiber crew and your asses up to the control room immediately".I had my fiber crew begin packing all of our equipment to relocate to the control room for the newest crisis, while I went ahead and went through security to scout ahead and try to expedite getting our equipment through the containment area's security check point, the X-ray and explosives detectors and so on.It takes my crew over an hour to clear security and haul all of our equipment up 6 flights of stairs, enter the control room and finally get the shift supervisor's attention from his recent bass boat purchase conversation and allow us to cross the “red line" to access the equipment room.While one of my team sets up the OTDR with a 1 kilometer test jumper cable box between the OTDR and the fiber panel (remember this for later please) the third tech and I calibrate our power meter and light source so Wade can go to the far end of the fiber cable to be ready for the actual loss testing.Trey and I quickly do our second OTDR shots over all 12 fibers and see none of the back reflections that Bernie claimed were showing up on his OTDR shots.As Trey and I complete the last of our OTDR shots, Wade calls in to say he is ready to perform the power meter testing and we complete the one way shots to him, then we have to meet Wade in the security checkpoint to swap the power meter and light source to shoot the fibers from the opposite direction.After all the testing has been completed, we compare the latest test results to our initial results from the previous day's testing and find no significant differences. As a matter of fact our power meter readings have improved because we had used new test jumpers that I kept in reserve for just such scenarios.I leave Trey and Wade to finish packing up the equipment and head to the Data Center with the new test results in hand.It's now after 9:00 PM and upon reaching my cubicle, who do I find sitting in my chair with his feet up on my desk, want to take a guess, yep, it’s none other than good old Bernie and his second in command Cody is there also.To say Bernie's attitude was smug, would be a huge understatement.We reviewed my printed out test results and Bernie is still insisting that there is a problem, so I finally ask to see their test results.Sure enough I see huge back reflection spikes on his printouts, but then I notice the footage markers where the back reflection anomalies are occurring and pull out my cable log that documents actual cable sheath footage between every termination and splice point for the entire cable length.The anomalies shown on Bernie's OTDR printouts are hundreds of feet off from our splices and they are repeating at the exact same distances for the length of the cable, but another thing I notice is that at each repetition, the back reflection is decreasing in a steady sequence.My mind flashes back to my Siecor classes and I realize that what Bernie is seeing on his OTDR shots is what is literally called a “ghost" in fiber optic terms.A ghost is caused by the connector being plugged into the OTDR having an almost perfect connection that creates a huge back reflection that the OTDR spuriously repeats at uniform distances with decreasing strength at each repetition.Remember what I asked you to remember earlier about my crew using the 1 kilometer test jumper box on our OTDR testing, that was used specifically to prevent ghosts by allowing the back reflection from the first connector to dissipate and allow the OTDR operator to zero in on the actual beginning termination without the false back reflections.I look up at Bernie from my cable log and ask him if he had bothered to look at his copy to compare the footage between the anomalies and the actual physical splices to make sure they were even isolating the correct cable segments for their expanded OTDR shots.Bernie's still smug reply was that he knew enough about OTDR testing that he didn't need to look at the cable log to find the splices, because of the back reflections.The other member of Bernie's crew Cody had spun the cable log around and was now checking the footage with a worried look on his face, when I asked Bernie if he remembered anything from his Siecor classes that talked about ………GHOSTS?Cody's immediate response was “OH SHIT" and I could see the color drain from his face, Cody had also attended the same Siecor classes with Bernie.Bernie's response was an immediate “BULLSHIT", so I walked over to my book case, pulled my Siecor manual from the class and quickly found the pages detailing what a ghost looked like with a very clear picture of a ghost. Then I laid Bernie's OTDR print out above the picture and it was almost identical.Cody repeated his earlier “OH SHIT" while Bernie's reply now came out as “OH FUCK YOU", then he started to argue again, that the splice protection sleeves were creating the problem.Cody is now on my side and tells Bernie, “we fucked up Bernie, Floyd's right".Bernie has finally lost his smug little smirk and his eyes are bugging out, the realization finally has set in enough that he even takes his feet off my desk.As Bernie and Cody start to walk out of my section's door, I calmly and maybe just a little smugly ask them if they want to tell the Dept Manager that they were wrong or if they would prefer me to explain the situation to him, Bernie mumbled over his shoulder that their manager would handle telling the dept manager the next morning, since we were apparently the only ones still on site.As I walked through our building's front door the next morning at 6:30 AM, I saw Bernie, Cody and their manager filing into the dept manager's office and just before their manager closed the door, he shot me a look over his shoulder similar to when a dog knows he has messed up bad.After dropping off my briefcase and lunchbox at my desk, I was in the breakroom pouring my coffee when my manager walked up to get his coffee and he told me to produce a report on the event, including wasted manhours and how much it affected us meeting the deadline for the other project that we were pulled off of. He had to speak louder than normal to be heard over the screaming coming from the dept manager's office.As I was tying up my report, the dept manager walked by my cubicle on the way to my manager's office, stuck his head in and said “BOO".I really expected Bernie to be terminated after this second false alarm of his, but I was wrong.You see there was another part to this story that I wasn’t aware of, do you recall earlier when I mentioned that my manager and the dept manager were both in the same Navy Reserve unit? Well it turns out that several other people in our department were also part of that unit including my friend Bernie. I don’t know if that was a part of the reason he wasn’t terminated then or not, but if you will bare with me a bit longer, we’re getting pretty close to the end of Bernie.After I completed my report on the previous day's circle jerk and took it to my manager's office, he went over it with me, suggested a few improvements, then instructed me to take it next door to a new member of the telecom dept that I had barely even met.This new guy was named Arnie and he was presented to our group as a “Cost and Scheduling Manager" that worked for Bechtel.We had been hearing rumors that Bechtel was being brought in to take over construction management to get the plant completed and on line, because the customer's own personnel were apparently incapable of getting the project completed.After I had made the revisions to my report, I delivered it to Arnie. Arnie remarked that he would take a look at it and would like to discuss it with me after he had time to read it.An hour later, Arnie paged me and asked me if I was where I could come back to the office and review my report with him.My response was that I really needed to stay at my current location, to help my fiber crew complete that particular task, but I should be available after lunch. Arnie then asked if we were by any chance testing with the OTDR and that if we were, he would really appreciate it if we could give him a demonstration.I told Arnie that I could have one of my guys that was picking up supplies swing by and pick him up to bring him to our location.Arnie spent the next couple of hours observing the crew testing the fibers with the OTDR and the power meter.It began dawning on me by several of Arnie's detailed questions that this was more than just a “dog and pony show”, but I didn’t have the experience or thought processes to see where this was all going yet.After a couple of hours my guys were wrapping up and Arnie requested that we return to our building so I could show him our documentation and processes for keeping it all updated.We spent almost another hour with me showing Arnie all of our test documentation and cabling logs. At the end of that, Arnie asked me who all were given copies of the information and I explained that Bernie was the only one that we gave a separate copy to in order to reduce the number of copies to update, that everyone else used my group’s master logs as they were not allowed to remove them from our area.Arnie thanked me for my time and I thought that was the end of it.A few more weeks go by and the Bechtel rumors prove to be true, the customer's dept manager is transferred to their corporate headquarters and Arnie is announced as our new dept manager.I'm still so busy trying to keep my engineering and fiber crews going that I've apparently been missing out on Bernie's latest claim that our new fiber installation is causing the plant's back up Microwave system to crash constantly.The first I hear of it is one morning when I walk in and Arnie calls me into his office, then asks me to bring my fiber cabling and testing logs with me.The first thing Arnie ask to be shown is the specific fibers that connect to the Microwave system's Rockwell DML-45.I flip to the proper OTDR and power meter readings and turn the logs to where Arnie can see them Arnie stares at the pages for a few seconds flipping back and forth between my crews initial tests results and then back to the final test results performed by Bernie and his crew. Suddenly Arnie puts his finger on the signature space and asks me if that's Bernie's signature and I confirm that it is as all of the final test results have to be signed off by Bernie, the same as all of my crew's test results are signed by me.Arnie thanks me for my time and doesn’t say another word, but I'm getting the sense that Bernie might have some “esplainin" to do, but I was still in the dark about Bernie's claims that the fiber was causing the microwave system to crash, but I did know from a previous staff meeting that Bernie was on vacation, so I'm thinking it's not that big of a deal for whatever Arnie was inquiring about.A few hours later and as I open the outer door on our building's foyer, I almost get knocked down by a large cardboard box coming through the foyer and don’t realize until the person carrying the box is past me that the person is Bernie.The situation still not dawning on me yet, I yell at Bernie's back, asking why he's on site, because he's supposed to be on vacation, all I get back is a “FUCK YOU ASSHOLE!” over his shoulder.Once I get to our engineering section, I ask one of my engineers named Jimmy “what's up with Bernie's attitude and why wasn’t he on vacation”?That's when I find out that after my earlier meeting with Arnie, he had called Bernie at home to request Bernie come in to the office for an urgent matter regarding the plant's microwave system.Bernie was expecting to come in to try and get the Microwave system back on line, but he got fired instead.Turns out that Bernie's sidekick Cody was pretty good at servicing Rockwell DML-45's and more knowledgable about testing fiber optic cabling to boot.[Big thanks to Scott Alexander for the many editing recommendations]

Twilio is a company founded by developers for developers! Twilio provides developer tools through PaaS model, or software-based platform which enables customers to easily add voice, messaging and video to their apps. Twilio is NOT selling a final product that can be consumed by an end-user (e.g., prepackaged software to solve their business problems), but rather is providing developers/customers with the prefabricated building blocks (i.e., APIs) needed to build any communication-based functionality they desire right into their application. By just virtually buying a phone number and swiping their credit card for potential per minute usage, developers can build contextually relevant communications by preventing their own customers’ users from leaving the application when they need to interact with someone (think why WeChat is popular in China). On this side of the world, popular use cases include ridesharing apps enabling anonymous communication between passengers and drivers and e-commerce companies sending automated delivery notifications or promotional messages.So how does it work?A developer signs up for Twilio, chooses a local virtual number (e.g., with 415 area-code for San Francisco, 212 for New York City) to send and receive voice, or SMS messages. The developer then maps the virtual number to a ‘request’ URL (the application’s URL, which Twilio would request from the developer’s application server when receiving a voice call on behalf of the developer/customer). The URL, or Uniform Resource Locator, created by the application developer, would describe to Twilio how to control the content of phone calls.The developer defines a set of business rules, or instructions to handling incoming and outgoing calls for each customer cohort. These instructions include: 1) Say - inform the customer that his or her order has been processed, or play a prerecorded sound file (message or music); 2) Gather - collect information from the caller; 3) Record - record the call; 4) Reject - hang up; and 5)Dial - dial this specific rep number to forward the customer to, or set up a conference call. In essence, on one end Twilio ‘ingests’ the phone call or the message flow, while on the other, it provides APIs for developers who, in turn, instructs Twilio on how to handle the incoming or outgoing phone calls. An API, or Application Programming Interface is a prefabricated block of software code that performs basic, reusable functions—e.g., displaying text on a computer screen, enabling inter-app communication—to allow developers to simply focus on building value-add, user facing applications.The call gets ‘load-balanced’ to a number of nodes/servers in Twilio’s cluster (managed by AWS). Each group of machines/servers perform specific roles—for example, some roles might be CPU-bound, a task is determined by the speed of the central processor, while others are memory-bound, determined by the amount of memory required to hold data—allowing Twilio to scale each group of machines independently of one another. The call comes in, gets load-balanced to a ‘Voice’ node cluster (i.e., a group of servers dedicated to voice function) to be answered by Asterisk, an open source PBX software platform.In general, a PBX allows telephone users to set up circuit-switched calls to other users in the same company (without toll charges) or to connect with users of the public telephone network. Effectively, PBXs shift some of the switching system out of the telephone company’s central office ‘CO’ and into the customer’s premise. The switch permits direct inward dialing ‘DID’ to a specific extension. Twilio stores the ‘DID’ numbers in its database; on the other end, the developer/customer would associate her or his assigned ‘DID’ number to a specific URL of a web application. As the call comes in, Twilio would make an HTTP POST request to that URL. In turn, Twilio would receive from the application server the XML instructions on how to handle the call.How does a phone call work? Historically, the telephone network has been a hierarchy of smaller networks performing different switching functions. In the US, switches were organized into classes from 1 to 5, Class 5 being assigned to the End Office or Local Exchange ‘LEX’. Classes 1 to 4 were Toll Centers or Transit Nodes ‘TN’. For a particular country, this network hierarchy would also be connected to international networks through a higher switch called International Gateway. Even if the modern structure has flattened, removing the Class 1 to 3 switches from the hierarchy, Class 4 and Class 5 designations remain in use today.The core of a telephone network is made of switches—TN and LEX. The primary function of these switches is to establish connections between telephones and the transmission equipment used to carry the voice call. When a called is placed, the switch detects that the receiver has been lifted, provides a dial tone and collects the called number. It then translates the dialled digits into a destination path by finding and reserving an idle circuit—if not it returns a busy-line tone. When the connection is established it sends a ringback tone to the caller until the destination picks up the phone. Then the connection is established by linking all the connections previously reserved in the transmission network and in the end terminates the call by releasing network resources reserved for the call.Typically, residential customers are connected to switches through an Access Node ‘AN’ or a Digital Loop Carrier ‘DLC’. This equipment converts the incoming voice into the appropriate digital format and also performs some basic call-processing functions. DLCs are access solutions that enhance the reach of the switch for remote customers. They do so by multiplexing many customer lines into just a few trunk lines that return to the switch.How does an internet connection work? The main historical way to access the Internet was a dial-up connection or narrowband access. Narrowband access uses the phone line to the Local Exchange, which is then connected to the data network through a Narrowband Access Switch ‘NAS’. The data network is typically a set of IP and/or ATM switches. The most widespread broadband internet access today is DSL. It re-uses the telephone copper wire up to Digital Subscriber Line access Multiplexer ‘DSLAM’ that directs the voice traffic onto the PSTN and the data traffic to the data network through a Broadband Access Router ‘BAS’. The PSTN, or the Public Service Telephone Network is a circuit-switched network that sets up dedicated voice circuits across a network of copper and fiber optic cabling.How do converged networks handle calls? There are many different ways to converge networks. In general, Network convergence requires the addition of two types of network elements: Media Gateways (MG) and Softswitches. Media gateways act as the physical bridge between the two networks. Their main purpose is to convert the TDM voice packet into data packets—generally IP or ATM. The softswitch acts as a media gateway controller by extending the voice network signalling ‘SS7’ into the data domain.When a phone call is placed on such a converged network, both the voice connection and the signalling are routed to the data network through a media gateway. The softswitch is a central database ensuring that Media Gateways properly communicate between each other.How does Twilio Fit in? To properly understand the inner workings of the Twilio platform, it is important to understand how basic web browsing work.What happens when you type a URL into a web browser. When you typeaws.amazon.com for example, into a web browser, a behind-the-scenes process occurs translating the domain name (aws.amazon.com) into the IP address (54.239.31.69). The IP address is sent to the web browser so the user can be connected to the website. This process, called resolution, relies on a global network of name servers. The browser is a program that requests information from a web application server, which in turn finds the requested information and sends it back to the user’s client (desktop, mobile).A web server is essentially a software program that sits in the middle tier of an ‘N-Tier’ computing environment between client-side environments (web browser), and server applications. Application servers allows developers to focus on creating specific business logic objects without having to code at the system level. The primary role of the application server is to access enterprise servers (e.g., AWS’ servers) for business logic and data, serving application objects to clients in the form of static or dynamic HTML pages, JavaScript, etc.Clients and servers communicate using Hypertext Transfer Protocol Language (HTTP). IP addresses are what machines on the Internet use to identify one another. It is a 32-bit address defined by Internet Protocol, or IP that uniquely identifies each computer on the Internet. Every computer, or machine on the Internet has a unique IP address. If a machine does not have an IP address, it is not really on the Internet. When searching for frequently visited Websites (e.g., aws.amazon, Facebook, Google, etc.), the recursive name server already has the information cached in its memory and passes the IP address on to the user’s computer browser without having to take any further steps.The ‘conversation’ taking place between the browsers and the local recursive name server is analogous to the following dialogue:User’s computer: “What is the IP address of the URL Amazon Web Services (AWS)recursive name server: “I already know that. The IP address is 54.239.31.69.” The local recursive name server then delivers the address to the end user’s browser.Under ‘the cover’, a ‘load balancer’ differentiates between a request to a main page and another to its extension (e.g., www.amazon.com and today’s deal extensionwww.amazon.com/gp/goldbox ),and sends the request for the web page of today’s deals to a group of servers that is optimized for that task. A load balancer goes one step further by allowing traffic routing decisions to be based on the so-called HTTP header (HTTP, or Hypertext Transfer Protocol is an application-level protocol that is the foundation for the World Wide Web). A load balancer’s controller not only examines the URL, but also information such as cookies, client source address, etcTwilio Client SDK enables VoIP communications. Twilio Client enables developers to integrate Voice Over IP, or VOIP into their applications.To illustrate, let’s examine a scenario whereby hypothetical InsuranceCo were to roll out Salesforce CRM Call Center mobile application through which is provisioned several customer service numbers (local numbers, international, 1-800 etc.) provided by Twilio. InsuranceCo buys a virtual 1-800 number from Twilio. InsuranceCo’s customer calls the 1-800 number. Salesforce has set up workflows, on behalf of InsuranceCo, for how to deal with inbound calls. Twilio Client SDK enables VoIP communications within the Salesforce application. The customer’s carrier (e.g, Verizon) routes the call to a Competitive Local Exchange Carrier ‘CLEC’ with whom Twilio has contracted (e.g., Level 3) . The call is routed through AWS to Twilio, triggering an HTTP request to Twilio’s call API to initiate the call.On the other end, Twilio 1) looks up the voice URI (Uniform Resource Identifiers) of the Salesforce Service Cloud application, and 2) provides Salesforce’s web server +information about the caller (e.g., originating and dialed phone numbers, date and time of the call, geographic location of the caller, etc). Twilio would also request instructions from the Salesforce application server on how to handle the call. The process is quite similar to how a browser makes a request to a web server discussed above.Once the Salesforce application receives the caller’s information (i.e., the call is logged into Salesforce). Processing function follows, involving a caller/customer database lookup to map customer’s caller ID to her or his historically stored information (e.g., caller's support ticket history). Caller’s information are then injected into the Salesforce application workflow. In this example, InsuranceCo.’s marketing database (maintained by Salesforce) has been capturing data about the customer’s behavior from various sales and communications channels.Twilio’s TaskRouter. Salesforce’s application server then sends to Twilio’s server (which is hosted on AWS) XML instructions on how to handle that specific phone call (e.g., since the analytical engine determined the caller is a high value distributor/customer, please forward immediately the call to highly skilled call-center rep A). An HTTP request is sent to a virtual call router’s API (i.e., Twilio’s TaskRouter's API). The call router, or TaskRouter authenticates, or looks up the account holder’s unique ID, matches the caller to the intended rep, then sends an HTTP request to Twilio’s server to initiate the phone call between the caller and the rep.The existing data on the caller/customer is enhanced by the overlay of the real-time contextual communication data elements; i.e., putting the customer call in context of the Salesforce’ customer service application, providing the call center agent real-time intelligence, needed for improved customer profiling. The agent can instantly access and view prior communications, use real-time analytics identify the personalized best next-action for the customer (e.g., promote a special offer, provide a service alert), and through, for example, Pitney Bowes’s content-creation platform, automatically integrates incoming data, compose new offer and send it to the customer through the relevant channel.What is Voice over IP ‘VoIP’? Voice over IP (VoIP) is a technology that involves sending telephone calls over data networks, such as the Internet, rather than the traditional Public Switched Telephone Network ‘PSTN’. Traditional phone calls across the PSTN use a dedicated circuit that transfers calls as uncut streams, allowing no other information on the circuit regardless of available bandwidth. Conversely, IP networks transfer data more efficiently in packets that get reassembled on the receiving end rather than using a dedicated circuit. For a VoIP call, audio is first converted from an analog signal to a digital signal through a codec, separated into discrete packets, sent across an IP-based network, put back into order at the termination point, and then converted back to an analog signal to create audio that the end user can recognize. There are three major VoIP protocol standards—namely, H.323, SIP, and MGCP.The Session Initiated Protocol, or SIP In March 1999, the Internet Engineering Task Force ‘IETF’ introduced Session Initiation Protocol ‘SIP’, which was designed to support quicker call set-up times and enhanced Web capabilities. SIP does not require gateways to maintain all call information during the life of the call and, as a result, it is less demanding of servers and is more scalable.SIP focuses on session initiation, modification and termination, and leaves the session and connection details to be negotiated by the end points. SIP uses a simple text command structure with HTTP syntax and URL addressing. Thus, SIP is well suited for any Internet- and Web-based applications.SIP allows direct communication between clients via a peer-to-peer ‘P2P’ connection using only IP addresses, and in most cases a SIP proxy will query a Domain Name Server ‘DNS’ to resolve a domain name into an IP address of record. An important characteristic of the SIP protocol is that it allows endpoints to establish a peer-to-peer connection and communicate directly. This is in contrast with other protocols that require resources on the network to control communications between end points.How does SIP “Work”? To initiate a phone call, the SIP phone issues an INVITE containing the caller’s IP address and the type of media requested (e.g., voice). From there, the SIP servers (Proxy and Redirect server) send back the SIP-URI where the called party can be reached. In this respect, the SIP servers somewhat resembles the DNS, or Domain Name Server servers on the Internet.What is the Domain Name System ‘DNS’? The Domain Name System (DNS), a global, distributed database infrastructure, is part of the fabric that holds together the Internet—performing the simple, straightforward function of mapping URLs (Uniform Resource Locator) to IP (Internet Protocol) addresses. Every Web server on the Internet has one or more unique IP addresses, represented as four sets of numbers, called octets, separated by periods (e.g.,140.211.169.9). The DNS enables people to use domain names (e.g., www.opendaylight.org ), which are much simpler to remember, to identify Web servers as opposed to IP addresses (e.g., 140.211.169.9). Each time a user enters a domain name into a computer’s Web browser, the DNS translates the user-friendly URL into the IP address needed to locate the appropriate Web server.SIP uses the above request and response method to establish communication among the various network components and ultimately enable a multimedia conference between two users. Users are identified by globally reachable unique addresses called URIs. URIs use the same format as email addresses to describe IP service points (e.g.,[email protected]). Users register their assigned URIs with the registrar server, which provides this information to the location server upon request. Users can have multiple URIs with different service providers that point to the same device, but they can also be reached with traditional telephone numbers. Calls using these traditional numbering schemes are translated into SIP URIs using the ENUM method.E.164 Number Mapping ‘ENUM’: Telephone Numbers on the Internet ENUM, or E.164 Number to URI Mapping translates telephone numbers into Internet addresses. ENUM is a protocol that merges the Public Switched Telephone Network ‘PSTN’ and the Internet—mapping complete international telephone numbers to URIs. Since the SIP protocol is IP based, it provides users (and applications) globally reachable addresses called URIs (Uniform Resource Identifiers). URIs use the same format as email addresses to describe IP service points (e.g., tel: [email protected], mms:[email protected], etc.) and can be reached with traditional telephone numbers. Calls using these traditional numbering schemes can be translated into SIP URIs using the ENUM methodology.To put ENUM into context with the aforementioned technologies, SIP performs the initiation of interactive communications sessions between users, as well as termination of those communications and modifications to those sessions. SIP is one protocol that may be used by ENUM to send out initiation attempts to multiple locations in order to find the user who is receiving a call. By placing telephone numbers into the DNS, ENUM can facilitate a range of applications including addressing for fax machines, email, instant messaging, etc.What Value Does ENUM Add? ENUM enables users to access Internet services from Internet enabled phones, ordinary phones that are connected to Internet gateways or proxy servers and/or other Internet devices that may have inputs limited to a numeric keypad.ENUM also provides users with greater control over communications, including allowing users to indicate their preferences for receiving communications. For example, a user can indicate a preference for voice mail messages over live calls during certain times, or may specify a call forwarding location.ENUM allows an end user to reach an IP device by dialing a telephone number rather than entering a URI. A traditional number is entered into the calling device, and the number is then transformed into a fully qualified address by an application or a device that supports ENUM.A developer’s customer dials a Twilio provisioned virtual phone number. The SIP proxy queries the ENUM DNS server to resolve the fully qualified domain name into a URI. The SIP proxy will then query a DNS server to determine the IP address to send the invite. To illustrate, let’s walk through a scenario whereby End User A dials a Twilio’s business customer, or User B:User A dials User B’s phone: +1-646-470-8021.Internet Gateway converts number to a domain name and queries VoIP local recursive name server: 1.2.0.8.0.7.4.6.4.6.1.e164.arpa. By using ENUM, e.164 numbers can be used in DNS by transforming the phone number into a hostname. This is simply done by reversing the numbers, separating each digit by a dot and then adding thee164.arpa suffix.13 For example, the number +1-646-470-8021 would be transformed to the fully qualified domain name 1.2.0.8.0.7.4.6.4.6.1.e164.arpa.Local recursive name server: “I don’t know that address, but I’ll check with a name server that does. Hold on for a millisecond.”e164.arpa Tier 0 server: “Here are the addresses for the authoritative name servers for the CC1 1.e164.arpa domain.”domain.com name server: “The URI for 1.2.0.8.0.7.4.6.4.6.1.e164.arpa [email protected].”User A’s telephone contacts User B’s (Twilio’s customer’s) telephone at returned IP Address. The local name server launches a query to the DNS, which responds with the IP address (e.g.,108.231.245.239) of the local proxy server associated [email protected]. The SIP proxy server in User A’s network contacts the SIP proxy server in User B’s customer network, and the proxy server in User B’s network then contacts User B’s destination SIP IP phone.When the called agent receives the INVITE request, it determines if it can accept the call. If it can, it will ring the phone and then send back a response to let the calling agent know that it is ringing. When the phone is answered, User B’s, or Twilio’s business customer’s phone sends an OK response back to the calling agent along with its media capabilities. Both agents agree on a media channel, and User A’s phone sends an ACK to User B’s phone.After responses and acknowledgments are exchanged, an RTP ‘Real-time Transport Protocol’ session is established between SIP IP phones of Users A and B, or the end-user and Twilio’s business customer.Traditional Voice and the Public Service Telephone Network ‘PSTN’. The voice telephone systems are referred to as the PSTN ‘Public-Switched Telephone Network or POTS ‘Plain Old Telephone System’. PSTN is a circuit-switched network that sets up dedicated voice circuits across a network of copper and fiber optic cabling. The structure of the traditional telecommunications industry varies by country and depends on the nature of the regulatory environment.In the United States, the industry has been pushed into a competitive model consisting of a variety of participants, primarily oriented around local exchange carriers ‘LECs’ that provide last-mile connection for consumers and businesses within specific geographies ‘LATAs’, and Interexchange Carriers ‘IXCs’ that provide long-distance services. A more graduated categorization includes ILECs ‘Incumbent Local Exchange Carriers such as SBC/Verizon), CLECs ‘Competitive Local Exchange Carriers’ (e.g., Level 3), IXCs (e.g., MCI/Verizon and AT&T), and ISPs ‘Internet Service Providers such as Earthlink and Prodigy/AT&T.Twilio interconnects to the so-called Tier 1 carriers - carriers that own or control sufficient portions of their underlying network infrastructure;e.g., Verizon, Level 3 - to provide the PSTN integration. It uses SIP (described above) origination and termination to ‘talk’ to originate and terminate calls on the PSTN.Traditional PBX. As discussed above, a PBX is a telephone switch located on the premises of a company. Traditional PBXs were typically hardware-based solutions that ‘sat’ inside customer premises ‘CPE’, providing businesses with the benefits of direct dialing, call forwarding, and a variety of enhanced services. Put another way, PBXs were originally designed for businesses as a cost-effective alternative to the provisioning of individual lines to each end-user from the phone company’s central office. The PBX is like a mini-CO, owned and operated by the corporation itself. In this respect, traditional PBXs reduced both line provisioning costs for the corporation and telecom services expenses associated with intra-office calls.Mostly Proprietary. Nonetheless, because PBXs are highly proprietary systems, enterprises have had to rely heavily on the PBX vendor to deploy or integrate any new applications. In a traditional PBX, there is a proprietary operating system running on a computer processor in a proprietary chassis. The applications are also proprietary, running on the same or separate processors. The interfaces—trunk cards and line cards— are also proprietary. In short, the traditional PBX is like a black box, with the vendor controlling virtually everything—the addition or adjustment of applications generally needs to be made by the PBX vendor. The proprietary nature of the technology is often predicated by its complexity. In fact, a high-end legacy PBX usually incorporates about 5 million lines of code.Limited Scalability. One of the most significant limitation of legacy PBX systems may be scalability. PBXs are typically designed for a specific number of users, and once the enterprise expands beyond that specific capacity, a new and bigger PBX needs to be installed. Sometimes, small businesses have to purchase a higher-end PBX than they need in case of possible future expansion, resulting in a particularly inefficient use of capital. It is also problematic to connect PBXs across multiple sites, and the signaling between PBXs is proprietary. Another key problem is that handsets which customers may have purchased for the lower-end PBXs often do not work with higher-end PBXs. As a result, customer upgrades to a higher-end PBX system often necessitate the additional cost of purchasing new handsets.Asterisk and IP-PBXs. Enterprise networking focuses on three primary goals: 1) Scalability, 2) Controlling the cost of communication through the most efficient use of technology and carrier services, and 3) Improving the productivity and performance of workers by distributing information to support their activities. With the advent of the Internet, IP PBX systems were introduced and allowed for phone calls to be placed over IP-based, rather than over TDM-based networks. In such an IP environment, distributed communications servers ‘IP-PBXs’ provide scalability and redundancy by sharing and quickly reconfiguring resources in the event of individual server failure. This redundancy and distributed processing is only feasible because the architecture separates the voice traffic from the PBX, leaving only call signaling and processing responsibilities to the PBX; hence, enabling independent scalability.Asterisk, on the other hand, is commonly used open source PBX software platform, developed in 1999 by a company called Linux Support Systems that later changed its name to Digium. The development of Asterisk is predicated on the idea that modularity, or separating a PBX system into interconnecting components—akin to a boxful of LEGO bricks—would enhance scalability. An Asterisk based IP-PBX is essentially a x86 communications server, running on Linux.Hosted PBX With Hosted PBX, PBX functionality is delivered as a service over the carrier’s network. Enterprise customers typically pay for the service under a leasing arrangement. Rather than having a PBX system located on the enterprise’s facilities, those functions are located in the carrier’s network and delivered over IP-based trunks to the enterprise.SIP Trunking involves the direct IP connection of a SIP-enabled IP-PBX and SIP-compliant VoIP service provider. It is an IP-based service provided by telecom operators (and Twilio) to connect an enterprise’s PBX with the service provider’s network. Put another way, SIP Trunking is an IP connectivity consisting of a single pipe which connects a service provider’s network to an enterprise IP PBX. As discussed earlier, SIP, or Session Initiation Protocol, is the protocol used to set up IP-based sessions between network endpoints such as end-user devices or servers. SIP trunks allow operators to provision VoIP voice sessions. Enterprises benefit from SIP trunking as it consolidates their voice and data networks, replacing premise-based connectivity, and thus reducing overall costs. In the past, enterprises had to connect bundles of physical wires—PSTN lines—to a business—PSTN endpoint. A SIP trunk eliminates the PSTN lines—reducing the number of SIP connections per port—and other associated equipment such as PSTN gateways.Thus, SIP Trunking offers a number of advantages over traditional TDM-based connectivity. First, SIP Trunking allows the enterprise to reduce its telecommunications costs. While many enterprises already save on the cost of voice calls between their sites by implementing IP-based PBX systems and using intra-corporation VoIP calling. Using SIP Trunking, enterprises can further expand their ROI by extending VoIP outside of the corporate LAN.The savings comes from: 1) Getting rid of traditional analog/POTS, ISDN BRI ‘Basic Rate Interface’, ISDN PRI ‘Primary Rate Interface’, or T1/T3 subscriptions; 2) More optimal use of SIP trunk bandwidth as both voice and data services can be delivered over the same connection; 3) Greater flexibility in purchasing voice capacity as enterprises don’t have to purchase lines in groups of 24 T1 or 30 E1 lines; 4) Flexible routing of calls to preferred carriers —long distance calls can be made for the cost of local calls; and 5) Lower operating costs of IP-PBX systems vis-à-vis traditional TDM-based PBXs.For most enterprises, the desire to save money is the primary force driving adoption of SIP Trunking. Least cost routing is an interesting example. Enterprises can utilize SIP trunks from multiple service providers and proactively route specific calls to certain carriers based on country codes—operators often charge different international rates based on availability, time zone differences, and geography.Moving beyond these types of cost-reduction initiatives, SIP Trunking enables a host of additional capabilities that enterprises can benefit from. As discussed earlier, the SIP protocol itself was designed to initiate all types of real-time communications over IP networks — not just voice. Today, enterprises are taking advantage of not only the significant cost savings afforded from SIP Trunking, but also the ability to improve enterprise productivity through the deployment of Unified Communications applications.Where Twilio fits in: Elastic SIP Trunking.Companies such as Acme Packet (acquired by Oracle) use their Session Border Controllers SBCs as SIP firewalls. A session border controller is a piece of network equipment or a collection of functions that control real-time session traffic at the signal, call control, and packet layers as they cross a packet-to-packet network border between networks or between network segments. SBCs are typically located at the perimeter of disparate IP networks, such as between headquarter and branch offices, and/or between call centers and enterprise data centers. They provide network operators with ‘policy-based control’ over VoIP sessions, furnish basic protocol inter-working and defend the carrier backbone against a variety of attacks.SIP trunks are VoIP trunks from service providers that use SIP for call control and routing,enabling enterprises to create a single, pure, IP connection to a carrier’s core network. This can be viewed as an Enterprise IP-PBX that peers with the service provider core SIP proxy. “Twilio-SIP is for use in TwiMl type applications to terminate or originate a call from a known SIP endpoint or address. Elastic SIP Trunking is utilized when you have an existing application or appliance that needs to have origination and termination capabilities (think a PBX like Asterisk/Freeswitch) and Twilio will be that provider”.Programming Messaging.Twilio’s Messaging API enables developers to embed text-based communications in their applications. Using the same virtual number, our hypothetical InsuranceCo (discussed above in the voice section) can use the same both make and receive voice calls, and send and receive SMS.SMS would follow a similar pattern of voice--the flow from caller/sender to receiver--yet with the addition of an SMS aggregator. Say for example an Uber rider contacts a driver. The Uber application would use Twilio’s API to generate and initiate the SMS to the driver’s number. The SMS is routed through AWS to a CELC (e.g., Level 3). The CELC would then route the message to an SMS aggregator with whom Twilio has contracted (e.g., Syniverse). The SMS aggregator routes the SMS to the driver’s carrier (e.g., AT&T). The driver’s carrier then routes the SMS to the driver.The SMS Aggregator. An SMS aggregator such as Syniverse (acquired by The Carlyle Group)is essentially a ‘clearinghouse’ provider that facilitates wireless roaming between different carriers’ networks. As a third party intermediary, it plays a necessary role in a complex telecommunications environment characterized by different network architectures, signaling standards, billing record formats, and network protocols. The aggregator serves as the middle hub connected to all carrier partners, allowing each to roam on other’s network (assuming a roaming agreement is in place).Continuing on with the Uber example, when the SMS is sent by the rider, a detail record is created. This detail record contains basic information about the SMS (e.g., who is the sender, where they sent, the length of the message, the carrier that authorized the message). This record is then stored in one of the several formats. For GSM, the format is known as TAP, or Transferred Account Procedure (TAP) file, while for CDMA the format is known as CIBER.The data record must now be communicated to the right partners—this is where an SMS aggregator comes in. Syniverse, for example, receives the information in its data center, aggregates the data, and distributes the information to the right carriers. The company then calculates the net obligation of each carrier based on the information detailed by the data records.The Short Message Service Centre (SMSC). SMS includes a number of distinct features, which I have highlighted below. These are made possible as messages are sent via an SMSC. The SMSC mainly acts in a similar way to a router of messages. However, it also acts as an important interface with other parts of the network and other systems on that network.In general, one of the main features of an SMSC is the ability to store and forward messages. If the receiving device is switched off, the central system stores the message until it receives a signal that the device is now switched on, when it will then deliver the message. In addition, the functions of an SMSC can include providing the senderInterface with Other Network Elements. In addition to the store and forward features, the SMSC can provide an important interface to an operator’s other applications and act as the router between these. For example, the SMSC may interact with the pre-paid billing system, location servers, user profiles and platforms for other SMS based applications.