Session Initiation Protocol – The Secrets

For multimedia communications, a standard has been around called the Session Initiation Protocol or SIP. My sister platform, The Museum of Telephony, has explained this in a way that it’s an app-driven telephony like interface.

Originally for the fusion instant messaging or IM, video and voice calls, SIP became an international standard for basic telephony extensions (or “stacks”) for such technologies to work over the Internet.  SIP is an open standard, in “theory”. As you read along, it’s going to become a cliche pretty quickly.

This isn’t telephony per se, and it gets extremely technical. And degrading men can act like fanboys of this technology too. Trigger warning! Link contains non-laymen content from an apparent misongyst!

I digress. Just bare with me.

While SIP is “open” in theory, the reason it being is because  systems administrators are too lazy to learn specific technologies they can’t Google. They call other people lazy but they never take personal accountability.  SIP is based on very specific, often open standards.  One is roughly about a dozen different commands to call the gateway or server or cloud similar to the traditional world where one goes off hook and makes a call. On top of this, there are codes for both the device or user when SIP calls can or cannot be placed, and these “response codes” are also “standard” for troubleshooting purposes.

SIP is strongest for bare communication purposes in the 21st Century, whether to connect trunks to phone systems, conference calling or basic two way calling.

  • Designed for mobility, literally meaning for locations where users their local Starbucks is the their primary “office” and refuses to “get a cube”; or relocate a telephone or to remove or add users.
  • Designed for users who want to basic two way calling, without additional lines, “appearances”, etc.
  • Designed for more “app” driven needs such a user who wants to check their social media status; or their Salesforce information.
  • Most importantly SIP as form of a “trunk” (that is if a telephone line is coming into a device like a PBX) is very flexible and enables consistent Calling Line ID or Caller ID or CLID as opposed to other VOIP protocols.

It can’t be underscored enough that SIP is very technical, but for this simple service requires a lot of planning, and strong engineering.

Can You Tell Me How to Get to SIP Street?

Because SIP was built with an intent to connect to other Internet Telephone networks, there is greater complexity. When I say “Internet” I am using a 1990s sense of the time when there were different flavors of servers, and “extensions” and “drivers” for other machines to talk to another. Remember there was AppleTalk, IPX, NETBIOS, DECnet, etc. While TCP/IP, the Transmission Control Protocol over Internet Protocol is the standard in 2018, I will continue to use “Internet” as an example because the following example is actually interconnecting logins, emails or telephones.



*fake number

SIP adds letters and numbers to contact a user whether its over the voice, IM or video unlike before it was just numbers. Also IP addresses and @ symbols  are also part of this new process.

The strongest practice is to use the middle example. In order to do this, you must have a functioning DNS server, where the phone service is registered, where phone service’s server is properly routing DNS requests, and most importantly if there is an LDAP Directory Server, that it would be configured to handle users with the middle example.

This is the path of how people are now wanting to be called by Name and not by Numbers. Consumer examples are Skype, Facebook Messenger, FaceTime, etc.

Connections to the Outside World

Several components for SIP connectivity are:

  • SIP Gateway: A router like device to listen to SIP traffic, and when it triggers a command from a user, to divert that traffic to landlines or POTS network. For instance, if a user dials 9 and 7 to 11 digits after 9 would then forward that “session” to a gateways to then convert to call outside. In a super-strict SIP world, this would only be used for “emergency dialing” to have reliable E911 service
  • SIP Proxy: Essentially a server that centralizes SIP calls and not overload the network with redundant SIP “requests” (which could be sending an IM or making a request to do a video call after a voice call has been placed.) Used especially at the last stop to the outside world as well. (This could also acting as part of the PBX.)
  • SIP Registrar: This service handles the extensions, it’s passcodes and/or devices to connect to external voice mail servers, or other phone systems, as those devices should be authenticated for security reasons.

SIP is also known to be both a singling protocol (meaning it triggers servers and devices telephony stuff) and both an application.

SIP vs H323 (Begin the Holy War!)

SIP is different from H323, an open Internet Protocol standard to stuff the traditional PBX telephony featuresets, most often proprietary signals. The advantage of H323 (while many vendors have been unfairly feeling the pinch to go to SIP for the “standards”

  • H323 despite how the Internet Protocol works, is a “hardwired” setup. In order for the devices to work, they must be on the same network, must be assigned to a specific IP address, of which is often the server or appliance. In traditional PBX setups or even Key Telephone Systems,  this was often the circuit board that the CPU resided, if the CPU was non existent, the phone could beep and ring if you poke at the buttons, trying to find it’s home. SIP is more “mobile” on the other hand where it’s designed to work outside the local network. The best practice to prevent toll fraud or getting hacked was having H323 systems tie remote users via VPN. Now because everyone wants to use the “Cloud” the best practice for clouds to work properly are domain names exposed to the outside network. To say domain servers are always reliable would be an understatement
  • It’s known as a Gatekeeper, as time, domain and other servers can be requested by a single IP address. In other telephony environments in recent years, the setup had to be a separate servers. One strong advantage to SIP is it can be resolved to a cluster of servers using a single domain name, sub domain, etc. This can make it more reliable.
  • The device and user are often synonymous to the device’s Media Access Control or Ethernet address, or their extension number. SIP is based on username and password, which can in theory work, but personalized settings are out of the window, but given the recent distaste this legacy technology by my generation, it’s more unneeded work by admins.
  • H323 is configured at the server level, that is providing the multimedia. If SIP is the option, there needs to be a complex of a set of separate servers for each of the services, and planning. Similar to the above bullet, but more instances, separate and desperate processes as well. This is more ciloed logistically despite it being “open”. It’s easy for the lazy admin who doesn’t want to RTFM a Mitel or Cisco documentation, because all he cares about is the ITEF standards or GTFO. As a result this type of reliability could be in question.


One thought on “Session Initiation Protocol – The Secrets

Comments are closed.