When Do We Start Slurping SIP?

Imagine if your Voice over IP (VoIP) (define) phone administration was as easy as using the Web. No more dropped connections, insecure sessions, lack of integration, or dependence on one vendor for systems. With Session Initiation Protocol (SIP), the long awaited promise of unified messaging may finally come true. No, SIP is not the latest in silly soft drinks; it is the latest emerging standard to address how to combine data, voice and mobility into one neat package. With its simple and integrated approach to session creation, SIP has the potential to transform how companies do business.

SIP is still very much an emerging standard with little concrete application and few commercial applications to show yet.

For the past few years VoIP has been quietly changing the delivery of telecommunications services. Most of the major telecoms have been upgrading their internal backbone networks to rely on IP (define) as a replacement for the older switching technologies. The administrative simplification and cost savings of merging data and communications networks into a seamless whole is compelling for any company.

Given all the great advantages, why hasn’t everyone switched to the new technology? The biggest reason why the convergence has not happened faster has been a lack of tools and standard protocols for establishing network connections. The result is that many existing market products use proprietary protocols to create the network sessions. This means that users must purchase all their equipment from one vendor for assured end-to-end connectivity; plus the systems tend to be single function, thus limited in their use.

The lack of a standard session initiation protocol has long been hampering the development of real unified messaging. So what is a good IP Telephony engineer to do? SIP is the IETF (define) community’s response to the problem of a mix of proprietary standards and no clear emerging winner among the vendors. SIP has been lurking under the radar for a few years, but now that it is finally crystallizing as a standard, it is time that it received wider recognition for its capabilities and potential to revolutionize interactive communications technologies.

Not a soft drink, a standard

The current official SIP version is RFC 3261, but there is a IETF working group actively working on an extended version per RFC 2543. According to the official IETF task force Web site, “SIP is a text-based protocol, similar to HTTP and SMTP, for initiating interactive communication sessions between users. Such sessions include voice, video, chat, interactive games, and virtual reality.” By only being concerned with session initiation and teardown, SIP must integrate with other widely available networking standards because it works on top or in conjunction with the other protocols.

What does that really mean?

When a person requests a Web page, he or she makes a request to the Web server, which then responses by distributing the information to the requesting computer. Because the Web server passively waits for requests, the amount of data transmitted is decidedly one-sided. Unlike the HTTP or SMTP standards which are primarily “pull” technologies, SIP is designed specifically for interactive or two-way network communication sessions. Whenever you use a VoIP phone, initiate a video conference, or download a streaming video, you need to create a “session”; that is your computer and the computer you are communicating with need to negotiate how they will establish the connection. When you initiate a two-way communications session, only once the information about session authentication and the type and speed of the connection are exchanged between clients can the real communications can start happening.

Essentially SIP is, at its heart, a set of standards for negotiating the session communications or framework themselves, not the contents. The IETF working group has defined the following model and architecture for SIP development:

  • Services and features are provided end-to-end whenever possible.
  • Extensions and new features must be generally applicable, and not only to a specific set of session types.
  • Simplicity is paramount.
  • Reuse of existing IP protocols and architectures, and integrating with other IP applications, is crucial.

The RFC 2543 working group is working on the following list of specifications that clarify and extend the existing standard.

  • bis: a draft standard version of SIP
  • callcontrol: Completion of the SIP call control specifications, which enables multiparty services, such as transfer and bridged sessions
  • callerpref: Completion of the SIP caller preferences extensions, which enables intelligent call routing services
  • mib: define a MIB for SIP nodes
  • precon: completion of the SIP extensions needed to assure satisfaction of external preconditions such as QoS establishment
  • state: completion of the SIP extensions needed to manage state within signaling, aka SIP “cookies”
  • priv: completion of SIP extensions for security and privacy.
  • security: assuring generally adequate security  and privacy mechanisms within SIP
  • provrel: completion of the SIP extensions needed for reliability of provisional messages
  • servfeat: completion of the SIP extensions needed for negotiation of server features
  • sesstimer: completion of the SIP Session Timer extension
  • events: completion of the SIP Events extensions (Subscribe/Notify)
  • security: Requirements for Privacy and Security.
  • compression: SIP mechanisms for negotiating and guidelines for using signaling compression as defined in ROHC.
  • content indirection: a Proposed Standard Mechanism to reference SIP content indirectly (by reference, for example using an external URI).

How will SIP benefit you?

Once SIP is widely adopted in the industry, new products will improve the security and integration of all two-way communications applications. With over 1,000 products and services listed on the industry trade group, SIP forum Web site, there certainly is much development activity already. The products range from the expected IP Telephony and call routing packages to a SIP NAT solution from snom, a German company. As their recent press release states, “The snom 4S NAT Filter is a software network component that enables SIP phones to make calls from private networks without support from the SIP user agent.” This is a major step forward, as SIP generally requires a proxy agent to work from a NAT network environment.

SIP is still a very new protocol. The SIP extensions outlined in RFC 2543 are only at the proposed standard stage. The working group has been actively working toward adopting it as a draft standard, the next step in the process. In September 2004, there will be a review of the working group status and a decision as to what further development work needs to be done. The outstanding issues continue to be the level of complexity and extension coverage without compromising security, bandwidth, and the SIP original charter.

What little software that has been developed using the new protocol is mostly highly experimental Linux-based open source code. The only end user software so far is limited to a few software phones packages, similar to the toll bypass Net2Phone system from a few years back. There are also a number of Web-based utilities to handle ID and session management (a big problem with any two-way communication technology). For the developers in the audience, there are plenty of opportunities for more applications.


Is the world ready for SIP? Without SIP, the development of unified messaging is that much more difficult, but SIP is still very much an emerging standard with little concrete application and few commercial applications to show yet. With the push to develop more VoIP, conferencing, and other session-based technologies, SIP will become increasingly important in the next year or two as it is incorporated in more commercially available products.


Latest Articles

Follow Us On Social Media

Explore More