Microsoft has made a preview version of Exchange 2016 available for download. The code is on track for final release later on this year as part of the Wave 16 set of Office server applications, probably in the late October timeframe, which would match the release cadence for the last few versions.
Microsoft says that Exchange 2016 is “born in the cloud”, a claim that should come as no surprise given that the vast bulk of Exchange development resources now work on technology for Exchange Online and other Office 365 components. Given that Exchange on-premises is approaching its 20th anniversary and the overall importance of the cloud to Microsoft’s long-term future, it is kind of inevitable that this release should be highly influenced if not dominated by what’s going on inside Office 365. I’ve already commented on this aspect twice, pointing out the kind of technology transfer that’s going on and concluding that Exchange 2016 is rather more like the latest service pack for Exchange 2013 than the kind of full-blooded brand-sparkling new versions we have seen in the past.
Still, it was interesting to see the line taken by the marketing whizz-kids in the EHLO post announcing Exchange 2016 preview and parse out what lies beneath the surface shine. The major points are:
Simplified architecture: This is a crusade that Microsoft has been on for some years now. Exchange 2013 is all about simplification. Exchange 2016 makes deployment even simpler by combining the mailbox and CAS roles. In other words, what has long been best practice is now embodied in the product, even if the folks at VMware don’t like this approach. Microsoft also says that coexistence with Exchange 2013 is easier, which is what you’d expect from a service pack!
Improved reliability: This is the area where a lot of technology transfer from the cloud is seen. If you were running hundreds of Database Availability Groups inside Exchange Online, you’d be rather keen that database failovers are faster (33% is the cited achievement). Also, the Replay Log Manager (available in all versions of Exchange 2013) is enabled by default to ensure that lagged copies don’t get into trouble if insufficient disk space is available. Microsoft uses lagged copies to protect databases in Exchange Online (Office 365 does not take regular backups, so the lagged copies protect against logical failures as opposed to the physical disk failures expected from the low-cost JBOD storage that they use).
More interesting is the intelligence being incorporated to detect when divergence occurs between database copies. This is an indication of lurking corruption that needs to be fixed before users complain of holes in mailboxes. Database copies that are going bad can be reseeded to be brought back to full health. Again, a great example of how to give software the ability to maintain itself.
Nicer OWA: Outlook remains the king-pin of clients when it comes to Exchange but Outlook Web App (OWA) is important too, especially in kiosk situations or when the only available network can’t handle the often insatiable demands that Outlook seems to make. So some new features come to OWA from the consumer Outlook.com (sweep), others (better performance and rendering) arise from engineering work, and others come from Exchange Online. The new interface is smarter and better to use than OWA in Exchange 2013, even if it now supports emojis. I guess that last feature is the must-have that will convince CIOs around the world to instantly make plans to deploy Exchange 2016…
Better Search: Microsoft has been beaten up regularly about the inconsistent search results delivered by Outlook and OWA. The solution is to make Outlook (but only 2016) search online whenever possible so that the two clients return the same results, which would be nice. As I reported from Ignite, “Microsoft has measured the success of user searches within Exchange Online and discovered that across 600 million searches, many use just one word to look for an item and one in eight searches return exactly nothing.” Hopefully the inline search suggestions supported by OWA will help some of those who find searching difficult. However, as these refinements don’t show up in Outlook, they might remain invisible to a large percentage of the installed base.
Data Loss Prevention (DLP): Introduced in Exchange 2013, DLP capabilities have been built out with additional definitions for sensitive data types and the introduction of document fingerprinting. Now we get 30 new sensitive data types for South America, Asia, and Europe. Some refinements are being made in transport for DLP rules too.
Scalable eDiscovery: Exchange 2013 switched over to the Search Foundation to share a common search engine with SharePoint. One of the side effects was that eDiscovery searches could only cover 5,000 mailboxes instead of the 25,000 (which could be increased) in Exchange 2010. The limit was lifted some time ago to 10,000 for Exchange Online and more work has been done since for the Office 365 Compliance Center and those improvements are now finding their way to the on-premises product.
Auto-expanding archives: As described in my archive on how expandable archives work, the idea is that archive mailboxes can grow in 50 GB chunks and be connected together in a chain so that a single logical structure is presented to clients. This feature is available now in Exchange Online and is another example of a transfer from the cloud. However, older clients have no knowledge of the mailbox chain and only know about the initial chunk, which is 100 GB in the case of Exchange 2016. So Outlook 2013 will only be able to see 100 GB of archive whereas Outlook 2016 and OWA can see the entire linked set. And of course, no mobile clients can see any archive data because these clients are concerned with today’s news rather than yesterday’s archive.
Hybrid improvements: Hybrid connectivity remains Microsoft’s unique differentiator in the battle for cloud email supremacy. Once connected with the Hybrid Connectivity Wizard, a single view is presented of on-premises and cloud components. At least, that’s the idea, and most of the time it works pretty well. And with Exchange 2016, Microsoft promises that Office 365 tenants will be able to keep mailboxes on-premises while making use of features like Office 365 Message Encryption and Advanced Threat Protection. Continual enhancement of hybrid connectivity is a good thing because it’s likely that Microsoft will introduce other cloud-based services in the future. Some of those services – or more correctly, probably all of those services – will never make it to on-premises servers, so being able to consume them by routing traffic via the cloud allows access where it would otherwise be impossible.
The preview build will change before final release. More code will come across from the cloud, tweaks will be made, and bugs will be fixed. It seems like Exchange 2016 will be reasonably easy to deploy for Exchange 2013 installations and moderately easy if you run Exchange 2010. That is, after you do the necessary up-front planning to master the level of detail that can’t be conveyed in an article like this. And ISVs have had the chance to upgrade their software so that you have the necessary backup, anti-malware, reporting, and other products necessary to flesh out the complete Exchange 2016 infrastructure. Based on previous experience, ISV updates will become available anything from a few months to six months after Exchange 2016 is formally released. However, the fact that Exchange 2016 is so close in terms of overall architecture to its immediate predecessor might accelerate this process.
The preview co-exists with Exchange 2010 SP3 RU10 and Exchange 2013 CU9, but not Exchange 2007. This is expected and is in line with Microsoft’s policy of supporting interoperability with the previous two versions.
The software is in good enough shape to conduct exhaustive tests but not to support production mailboxes (except your own, of course). And the Edge transport is available and even some documentation is available. What more could anyone want?
Follow Tony @12Knocksinna