The U.K. is Microsoft’s largest subsidiary and has always been an important and mature market for Exchange Server. The results of a survey about the pace of transition to Exchange 2010 reported by the well-respected Computing magazine are therefore interesting and deserve some commentary.
According to Computing, Exchange has 80% of the enterprise groupware market with competitors such as Lotus Domino limping on with just a few percentage points, probably in companies that have used Lotus Notes since the dawn of time. Computing did two surveys in August 2010 and February 2012 to assess how quickly enterprise customers were deploying Exchange 2010 and reports that with Exchange 2013 on the horizon: “history suggests that whatever the advances in functionality may be, there will be no rush to adopt by UK enterprise.”
They point to the fact that the February 2012 figures (some 30 months after the original release of Exchange 2010) show that only 42% of enterprise customers now run Exchange 2010. And although there’s been a marked increase from the 8% reported some 18 months previously, 56% of customers still run either Exchange 2003 or Exchange 2007 with another 2% stuck on Exchange 2000.
My mind still can’t quite get around the fact that 2% of UK enterprise customers run Exchange 2000 as I wonder why they have never upgraded. The base survey data is from “IT Decision Makers at large UK organizations”, so I assume that these individuals are visited regularly by Microsoft salespeople, whose scorecards must look pretty bad with Exchange 2000 still in the picture. Couldn’t these accounts have been persuaded to upgrade to at least Exchange 2003 rather than stay on software coded in the last millennium? In any case, the 2% has remained stable between the surveys so these customers must be set in their ways.
The report observes that many customers who run Exchange 2003 or Exchange 2007 are now “en route to upgrading to what is – numerous bug fixes and service packs later – a stable and mature platform.” The platform here refers to Exchange 2010, even if Computing is incorrect in some of the advantages listed for Exchange 2010, such as support for Outlook 2010 and virtualization, both of which can be gained with Exchange 2007. That being said, there’s no doubt in my mind that Exchange 2010 is a far superior product to Exchange 2007, if only because of its high availability features.
No mention is made about a potential shift away from Exchange 2010 as the target destination to either hybrid deployments or an “all-in” migration to Office 365. Perhaps this was due to the date of the survey as Office365 was still under the cloud (no pun intended) of some early reliability hiccups in February 2012 and Exchange 2010 SP2’s hybrid connectivity wizard hadn’t made an impact then.
I wonder if the advent of Exchange 2013 after its debut at the Microsoft Exchange Conference (MEC) in September will impact upgrade plans further, especially if, as has been the practice in previous versions, Exchange 2013 requires a client upgrade (Outlook 2013) to achieve its full potential or some features won’t be quite complete until a future service pack arrives. The fact that Outlook 2013, along with its Office 2013 counterparts, boasts a new user interface is likely to cause further delay due to the impact of any UI change on users and help desks. We shall see in due course.
The report noted that the main reasons for delaying a move to a new release were due to cost (36%) or fear of disruption to end users (27%). The need to deploy new or upgraded hardware is also a factor, as per the requirement to deploy Exchange 2010 on Windows 2008 R2 and to upgrade Client Access Servers (CAS) to take on the increased load generated by moving the MAPI endpoint for clients from mailbox servers to the CAS. Exchange 2013 is likely to cause some similar debates around choice of operating system (will Windows 8 Server be in the picture?), the kind of hardware platform used (dedicated servers or virtualization), and indeed the time required to deploy and then move user mailboxes from old servers to databases on new servers.
Interestingly, 34% of those surveyed reported that email performance had suffered some degradation in the immediate aftermath of the move to Exchange 2010. Thankfully this figure decreased over time to 3% at the time of survey. I think this indicates the usual experience where administrators grapple with new features in the early days of deployment and gradually become more experienced and competent over time. It might also indicate common problems with new software such as problems with third-party packages that interface to Exchange (backup products are a usual suspect here). Good planning and practice in a realistic test environment can reduce the potential for problems for the actual deployment but never eliminate snafus. It’s just the way that IT works.
It would be interesting to compare and contrast the data for the Exchange installed base in other geographies. Alas, the data isn’t available – at least, not publicly (I’m sure Microsoft has it).
Follow Tony @12Knocksinna