More fearless predictions

In our last newsletter we predicted a quiet year for network service providers, an uncertain year for WAFS, and a dynamic year for application optimization. In this newsletter we will look at Web services and MPLS.

In our last newsletter we predicted a quiet year for network service providers, an uncertain year for wide-area file services, and a dynamic year for application optimization. In this newsletter we will look at Web services and MPLS.

In 2005 we did a lot of research into the use of Web services and ended up fully convinced that the use of Web services for both application development and application integration has gone mainstream. In 2006 we are going to begin to see the impact of Web service deployment.

For example, we’ll see an impact from the fact that two of the primary protocols that enable Web services – XML and Simple Object Access Protocol (SOAP) – are heavy protocols, meaning that they place a lot of traffic on the network.

Another effect will be additional pressure for application optimization. The reason is that it appears as if it will be common to deploy a Web services-based application in a manner that has the Web services running in separate data centers. This means that even though the Web services may make frequent communications with each other over a WAN, the overall application still has to perform as if each component were running over a LAN.

For the last few years the trade magazines have been hyping MPLS. However, the number of articles devoted to MPLS has been sharply reduced in the last few months. Typically when the trade magazines stop writing about a technology that means one of two things: that either the technology has died, or just the opposite – that we are now actively deploying the technology.

It is clear that in 2006 we will see numerous deployments of MPLS, the vast majority of which involve implementing an MPLS service from a service provider.

Everyone we have talked to that has deployed MPLS has complained of the complexity associated with evaluating and managing MPLS services. This complexity comes from a number of sources, including the need to understand the differences in the various service provider offerings, whether it is how they manage QoS end-to-end or the various ways they price their services. The complexity also comes from trying to assign traffic to classes of service and then managing those classes.

Regrettably, the complexity associated with MPLS is not going away. If anything, based on the service provider consolidation we mentioned in the last newsletter, it may well increase. We predict that 2006 will be a very busy year for those IT organizations that are implementing MPLS.