Right now much of the conversation surrounding mesh computing seems to be dominated by Microsoft's view of the world given the fact that synchronization is at the heart of its Software Plus Services strategy.
But the problem with this viewpoint is that it assumes that mesh computing is going to be all about the synchronization of client applications running on smart phones and PCs with Web applications. While that's an important piece of the overall mesh computing equation, it seems to be a view of mesh computing that is fixated on where we are now in terms of computing versus where we are likely to be in two years.
As smart phones get more powerful, it seems that the need for synchronization between the client and the Web application would be greatly reduced. We are already seeing SAP talking about running the client portion of its CRM application natively on a Blackberry from Research in Motion (RIM). As smart phones increasingly become the most prevalent computing device used by people the providers of applications are going to find that they will need to run the client side of their applications natively on those devices.
As that begins to happen the data that is entered into those devices is automatically fed back to the Web application. There is no separate mobile environment that needs to be synchronized with the Web applications running on back end servers.
None of this says that mesh computing won't be important. In fact, Marc Canter has laid out a case for why we need an open approach to mesh computing that makes a lot of compelling points, most of which are focused on the back-end services that will be need to let people create composite applications, also known as mashups, on the fly regardless if their data is in an enterprise application or some service in the cloud run by Google, Amazon, Microsoft or whoever.
Right now Microsoft is trying to step into a void by saying in the absence of any widely supported approach to mesh computing it is volunteering to provide the glue as a proprietary service. While that may be a valuable service in the short term, long term it is probably not in the best interests of an industry where the whole is always worth more than the some of the parts thanks to open interfaces.
What all this implies is that the various industry titans that have an interest in mesh computing need to come together in either an ad hoc fashion or through some industry standards body to create an open framework for mesh computing that will not only drive the next generation of industry growth but also better serve users by giving them the underpinnings of open infrastructure for managing their data in a way that could provide the biggest boost to productivity since the creation of the spreadsheet.