Why I don't think on-premises Exchange is dead

One of the questions raised by some attendees at the recent Exchange Connections conference is the future of on-premises software. Apparently there’s quite a few people who think that Microsoft us busily killing off all its on-premises editions of server applications in the rush to embrace the cloud. In other words, products like Exchange 2013 might be the last in the line because Microsoft wants customers to use Exchange Online in Office 365.

I don’t buy this line of argument for many reasons. Here’s some of the more obvious that come to mind.

First, Microsoft would be silly to mess with the large and profitable Exchange installed base that they’ve been building since 1996. Some of the base will move to the cloud, but if companies lose faith in Microsoft, they might look at the Google alternative. I sincerely doubt that Microsoft would deliberately open up their flank to an attack by Google.

Second, once customers begin to lose faith in Microsoft’s commitment to on-premises software, Microsoft’s ability to sell other products into the Exchange installed base like SharePoint, Lync, Project, upgrades to Windows and Office, and so on is diminished. Account control is weakened further by doubts in customer minds, which is never a good situation for a vendor.

Third, although there’s no doubt that a large engineering group is doing lots of work on Exchange Online and less on Exchange on-premises, you have to remember that such a division of labor is unsurprising given the relative point in the life cycle of the two platforms. Exchange on-premises is very mature and feature rich and therefore needs less engineering (but it needs to be at the right level of quality). On the other hand, it’s a huge engineering task to create the kind of software that functions effectively and efficiently in a massive multi-tenant environment given the type of functionality that Office 365 delivers.

Remember that Microsoft now has a single code base for both the on-premises and cloud versions of Exchange, which means that developments from the service make their way to on-premises customers.  Managed Availability is one obvious example in Exchange 2013. Another can be found in the platform improvements such as making the CAS a stateless server and removing the need for load balancers to manage affinity for web connections. Sure, Microsoft needs this kind of improvement to make Office 365 work better, but the work also benefits on-premises customers.

It’s true that Office 365 gets a lot of attention these days and that it sometimes seems that the Microsoft sales force wants to sell nothing more than a few more cloud seats. It’s also true that using Office 365 is the right and most appropriate future direction for many customers, especially those who have less than 1,000 users and who struggle with some of the challenges to maintain, secure, and upgrade their IT environments.

Metrics drives behavior and I suspect that some sales incentives are responsible for some of the activity and buzz that we see around Office 365. However, as pointed out above, there are many reasons for Microsoft sales teams to keep their on-premises customers happy.

The number of people running the on-premises version of Exchange will reduce over time as Office 365 takes on more of the load, but for the reasons outlined above, I feel pretty confident that we’re going to see Exchange 16 in the next couple of years (Exchange 2013 is Exchange 15, not the 15^th version of the product but rather part of the Wave 15 releases of the Office applications). And Exchange 17 will follow Exchange 16 as sure as night follows day. Would anyone bet against this happening?

Follow Tony @12Knocksinna