January 2004

SIP: Redefining IP Voice Communications

BY DR. HENRY SINNREICH

Global IP communications over the Internet is probably the most significant development since the emergence of telephone networks over 120 years ago. New communication services for text, voice, and video, as well as mobility and presence-based services, enabled by the Internet, are leveraging a relatively new protocol known as SIP, or Session Initiation Protocol. SIP is an IP-based communications protocol that provides a network-based convergence solution. SIP is an open, non-proprietary standard modeled after HTTP that seamlessly integrates with the Internet. Because of its easy integration and ability to inter-operate with other protocols and vendors, SIP is the foundation for flexible and scalable VoIP solutions. This article will focus only on voice, but we suggest you keep in mind the other IP communication services enabled by SIP as well.

Present telecommunication networks and services provide adequate and universal telephone service at affordable rates. So, why would enterprises want to transition from their traditional circuit-switched networks to VoIP? Good question. An enterprise has two major incentives to complement existing services or migrate to IP communications. The first incentive is the potential short-term savings by moving to end-to-end VoIP and mixed IP-PSTN and also by moving PBX voice to VoIP usage. The second incentive, and motivation for this paper, is increasing overall productivity through the integration of voice, data, and productivity applications. Examples of integrated IP communications based on SIP (presence, instant messaging, voice, video and application sharing), with applications are Microsoft Office 2003 and other similar or related products, such as Siemens Openscape.

Existing VoIP Protocols and SIP
A large number of VoIP protocols are in use at present, reflecting their origin or the voice strategy of particular vendors. H.323 was introduced by the ITU-T in the late 1990s, and it complements H.320, H.321, H.322, and H.324, which are for audio/video conferencing using analog, ISDN, ATM, and the public telephony networks, respectively. H.323 was thus mainly modeled by ISDN and modified to work on IP LANs. It was used later by vendors for IP LANs and WANs.

The circuit switch vendors applied the central control model as used in PBXs for servers to control desktop phones over the IP LANs and the result were several master-slave protocols, such as MGCP and MEGACO/H.248. The central controllers used by such master-slave protocols are also referred to as �softswitches.�

In contrast, SIP is a peer-to-peer protocol, while the softswitch protocols and even H.323 presume some central control. SIP can use servers for network-based services, but the user can choose whether to use them, similar to pointing a browser to any Web server of choice for the home page.

As for the many vendor proprietary protocols, such as used by IPBX vendors, they carry all the disadvantages of vendor lock-in of the customers into proprietary servers and IP phones, for which in most instances there is no benefit from competition.

Distributed Networking Applied to Telephony Services
The architectural principles of the Internet -- avoiding single points of failure, avoiding single paths, avoiding state in the network, and end-to-end control -- have proven to support more reliable communications than circuit-switched networks, in spite of their �carrier strength� network elements. Internet service has degraded somewhat with recent widespread virus attacks, but has never failed and has been proven so far to meet design expectations.

SIP-enabled networks are also distributed by design and have most state and control pushed to the edge of the network, in compliance with Internet architecture principles. Internet resilience is based on IP endpoints -- autonomous hosts that communicate over any available path and do not depend on the network for state. This is a fundamental difference from PSTN systems where a central authority activates and shuts down all processes in various boxes having a stake in the communication and where all systems in a path have to keep state for every call going through.

Higher-than-PSTN resilience for VoIP can be obtained if the underlying customer network is multi-homed to the Internet using truly diverse geographical paths and by deploying distributed SIP servers and distributed functionality for all SIP components, such as application servers, media servers, SIP-PSTN gateways, etc. This may not be the case however in many deployments, where lowest cost is the dominant factor. Events such as the 9/11 attacks in New York and the East Coast blackouts have confirmed however that VoIP deployed at the edge, is more resilient than the telecom infrastructure -- provided there is backup power on the site.

How SIP Improves Upon PSTN Functionality
Aside from the significant voice traffic migration to mobile telephony, the advent of the Internet and its associated technologies are forcing telecommunication companies and their vendors to face the facts that telecommunication networks have a competitor in the Internet and the deployed PSTN technology is obsolete.
In the context of VoIP, a closed system cannot be expanded with components from competing vendors using public standards. Examples of closed systems are softswitches and proprietary IP PBXs. In such systems, it is not possible to choose the following from different sources: IP phones; IP-PSTN gateways; service controllers; SIP servers; IVRs, VoiceXML and speech recognition systems; media and announcement servers; conference bridges; voice mail and unified messaging; and other components.

SIP, however, brings choice and several enhancements to traditional telephony; including call transfer in voice mode, automatic callback feature, multiparty conference calls, Web-sharing and data collaboration, and mobility. Visual voice mail alerts also falls into this category. The technology transforms e-mail messages into an MP3 or .Wav file, which can then be �viewed� (listened to) as an e-mail attachment on a secure Web site.

Innovative IP Communications Based on SIP
Using the development of the Internet communication protocols and SIP, IP communications enable radically new capabilities and applications. Probably the most far-reaching disruptive engineering decision of IP communications is to integrate voice and all other media (text, voice, video games, etc) with the Web with regards to addressing, protocols, and data formats.

SIP-based communications functionality will drive improved productivity by making it easier and more efficient for end-users to communicate. At the same time, SIP-based communications enhance flexibility by providing enterprises a variety of platforms, endpoints, and servers to utilize.

The Internet development community, for example, prefers text-based messages for easy code development and debugging. SIP uses text-based messages for simplicity and easy troubleshooting. In addition to working with standard telephone numbers and extensions, SIP can also use e-mail-like addresses. These addresses can be imbedded in documents and Web pages for �click-to-dial� applications. For example, SIP addresses can look like sip:[email protected] or sip:[email protected].

Presence-Based Communications and Instant Messaging
Presence will be a major new communications capability that is not available in circuit-switched telecom networks. A simple example of �presence� is an instant message (IM) client�s �buddy list,� which lists the user�s �buddies� and their current state -- online or offline. Additional state information, such as whether they are currently active or idle and whether they are currently typing a message response or not, is also provided using presence.

Presence can be used for such services as:

� Make �polite� calls, only when you see an encouraging or �smiley� icon;
� Avoid phone tag during busy hours;
� Automatic call-back on presence;
� Ad-hoc conference calls based on presence;
� Avoid waiting for call center agents -- replace ACD (Automatic Call Distributor) with agent presence showing also the length of the queue;
� On-the-air presence for mobile phones;
� Presence coupled with location.

The following scenario would only occur in a SIP-enabled environment: A mobile user or a user on the PSTN is trying to call a user that has an IP communication service and can be reached at a SIP phone and/or IP devices. The call is first routed by the PSTN gateway to the SIP server for the called party and from there to the SIP phone. If the called party does not answer, the SIP servers will proxy the call to the unified messaging (UM) server. The caller can now leave a voice message. The notification for �Message Waiting� will appear as a flashing light on the SIP phone (as is usual on PBX and mobile phones), as an e-mail notification and also on a UM Web page.

Making a VoIP Investment
One U.S.-based company that benefited from SIP-based communications, specifically VoIP, is Storage Area Networks, Inc. (SANZ).

SANZ was finding that operating separate data and voice platforms was costing a great amount of time and money as newly acquired companies had to be added to their enterprise network. The company also spent a considerable amount on long-distance charges between satellite offices. Given that, SANZ searched for an alternative technology solution that would reduce costs, simplify voice and network services deployment and support the company�s technology needs.

To face its challenges, SANZ selected a network-based IP communications service to streamline their needs over a single data network. In doing so, the company consolidated local, long-distance, data, and other network services through just one connection. They quickly found that a significant advantage of VoIP is the ability to avoid the toll charges of ordinary telephone service; In addition, the technology prioritizes voice packets over data, ensuring optimal quality of service.

With that challenge surmounted, SANZ is now focusing on what it really wants to do, that is cultivating its business through strategic acquisitions.

Conclusion
As voice is now viable over IP, the advantages of using the global IP network for converged voice and data services are just beginning to reach the marketplace. Furthermore, responsible IT managers will invest in new services that reduce high operational costs associated with existing services, such as PSTN/PBX telephony. By deploying IP voice communications and rich end-to-end controlled IP, IT managers will quickly realize implementation cost is negligible compared to the huge investments and operational costs required by traditional PBXs, by non-standards-based (not SIP compliant) IPBXs and by PSTNs. Last but not least, there is a wide selection of standards compliant vendor equipments to choose from.

Dr. Henry Sinnreich is a Distinguished Member of Engineering at MCI where his responsibilities include the architecture of IP communication services, SIP telephony devices, wireless, mobility for IP communications and quality of service as well as network support for communications. MCI owns, operates, monitors, and maintains one of the largest communications networks in the world. The company�s network facilities span the globe in more than 125 countries and over 2,800 cities.

If you are interested in purchasing reprints of this article (in either print or HTML format), please visit Reprint Management Services online or contact a representative via e-mail at [email protected] or by phone at 800-290-5460.

[ Return To The January 2004 Table Of Contents ]