Warning: SIP is no savior

Beware of a growing threat to the orderly management of enterprise communications.

There is a new school of thought that turns a traditional argument on its head. The old view is that open standards permit competition and enable enterprises to adopt a multi-vendor strategy without technical complexity. There is a new view that certain open standards may be too complex to develop.

Therefore, a multi-vendor strategy is more easily achieved by allowing multiple solutions. However, much depends upon what type of solution we are discussing. Let’s consider session initiation protocol (SIP), which has been the great hope for a universal call-control protocol in IP telephony. SIP is suffering in both the enterprise and public services domains. In the former it is recognized by most vendors as the default protocol, yet plug-and-play interoperability remains a long way off.

Dominant vendors do not want to reduce their products to commodity status, much preferring to make it uneconomical for enterprises to migrate to other vendors. Combined with the growing complexity of accommodating public networking and messaging, SIP might collapse under it own weight.

With public services, such as instant messaging (IM), most providers have ignored SIP. Even the International Telecommunications Union (ITU) is concerned that SIP may not be the best approach; accordingly, it has initiated a study into a potential replacement (H.325). This gives a further boost to those who dismiss such standards.

The primary argument is: what’s wrong with downloading multiple clients? This is what people do today if, for example, they want to use multiple IM systems. Interoperability can be achieved without each of them using the same underlying standard. This view turns toleration of multiple solutions into promotion of them. It also ignores the current dependency on the public switched telephone network, which acts as a lowest common denominator for certain signalling and addressing.

From the desktop perspective, a multi-client environment may present an increasing management problem for the enterprise. From the mobile perspective it’s a potential nightmare. We already face the prospect of tri-mode, (cellular/Wi-Fi/WiMAX) handsets. Imagine also an environment with no common public signalling or addressing.

The handset resources (memory, processor, power, etc.) required to accommodate a multi-everything environment would be unreasonable; interoperability would be at the whim of competing vendors.

It has usually been within the control of the enterprise whether to pursue a single- or multi-vendor strategy — but not so for network services.

An enterprise could select a solution for its own employees, only to discover that it requires a multi-client strategy to communicate with customers and suppliers that do not use the same solution.

In some cases, service providers are tackling this issue through bi-lateral arrangements. That does not reduce the need for enterprises to raise their awareness of this potentially growing problem and to develop appropriate strategies.