In case you missed it, Microsoft’s Ignite 2017 Conference was held last three months ago in Orlando, Florida, and it became very clear there that Microsoft understands that not every customer is ready or willing to “go to the cloud” because Exchange 2019 will be released to customers who want to run the next version of Exchange in their on-premises environments. While we are still probably almost a year away from a Customer Technology Preview of Exchange 2019, here’s what we can tell you now about the tenth version of Exchange.

What was announced

The next version of Exchange on-premises will be Exchange 2019. Of course, you already got that bit. Here’s what else was announced. Many of the innovation born in the cloud with Exchange Online will be included in Exchange 2019. Just what those may be is not yet public, but they were very clear in their hints dropped that security, manageability, compliance, and usability from Exchange Online will all be present in Exchange 2019.

What you can infer

While Exchange 2019 is going to be a new on-premises release, it’s unlikely that it will be an earthshattering change in how Exchange is done, so don’t think you will have to redo your entire on-premises environment to support it. What you should expect though, based on hints at various Ignite sessions, the latest requirements in Exchange 2016 CU7, the Exchange 2016 PLA, and what new requirements/compatibility issues that came up when Exchange 2016 was released all go into the following expectations/predictions.

  1. AD forest and domain functional levels will need to be at least at Windows Server 2008 R2 or higher. I won’t be surprised if, at RTM, this is actually going to be 2012 R2, but we’ll see.
  2. There will be no backwards compatibility with Exchange 2010, let alone 2007. Since 2016 would not coexist with the at the time N-3 2007, I expect that 2019 will not coexist with what will be the N-3 2010.
  3. To be compatible with 2013 and even 2016, those Exchange platforms will have to be at the latest CU versions available at the time…not the N-1 versions.
  4. Yes, this version of Exchange is for the on-prem environment, but will have even easier ways to establish hybrid coexistence with Exchange Online, and will make it easy to use Exchange Online Protection and Advanced Threat Protection for messaging hygiene.
  5. This is very much a prediction, so take it with a grain of salt, but to that end, there will not be an Edge Transport role in Exchange 2019.
  6. There will be one role, the Exchange server. This is working really well with Exchange 2016 deployments and, with Office 365 as the proving ground, obviously scales to any number of users you might need to support, so will continue in Exchange 2019. And if I am right that there will be no Edge Transport role, there will be no need to call out a server as a Mailbox server, so it will just be an Exchange server.
  7. Another prediction; the minimum Outlook client version supported will be Outlook 2010. However, Outlook 2010 will have less than a year of extended support remaining and will have to be fully patched to work. There will be no Outlook Anywhere in Exchange 2019 and if you are considering Exchange 2019, I don’t think you should be considering Outlook 2010 at all.
  8. Office 365 is managed with what is usually referred to as “zero standing access.” RBAC, combined with Just in Time (JIT) administration, greatly reduces the level of rights to the system at any point in time. Here’s another prediction-this will be in Exchange 2019 out of the box.
  9. My final prediction is a bit fatalist. Public Folders will still be supported. My advice is not to mention that and try to finally get rid of those Public Folders once and for all!


Modern Authentication, which is used extensively in Office 365, will also be available as an authentication mode in Exchange 2019. Based on the talk from Greg Taylor, Modern authentication for Exchange Server on-premises, this may be something you want to consider, especially if you are already doing some workload(s) in Office 365 or Azure. This may be very appealing to customers who want to keep Exchange (or some part of their users) on-prem, but still take advantage of things you can do in Exchange Online, like Conditional Access. It will also prove useful for customers who need to support OAuth, like universities and managed service providers.


According to a very short blog post from the Exchange Team, the plan is to share more details about Exchange 2019 next year, with a preview release planned for the middle of 2018, and a release date before the end of 2018. So, we’re about a year away from Exchange 2019 RTM if timelines hold. That means that Exchange 2013 is a year away from being considered N-2. Of course, the more important thing to note is that it will entered the extended support phase on 2018-04-10 regardless of whether or not Exchange 2019 ships before the end of 2018. If Exchange 2019 does release as planned, it will continue the trend of naming Exchange releases like car models-the year of release is before the year in the name.


One significant bit of detail that came out of Ignite regarding Exchange 2019 is that Microsoft is using solid-state drives to store indexes and more for Exchange mailbox databases. Whether Exchange 2019 will have this as an optimization, make it a recommendation, or determine it is a requirement is yet to be determined, but it’s a significant shift from the “Just a Bunch Of Disks (JBOD)” approach in previous versions. Called the MetaCache DataBase (MCDB,) SSDs will store indexes, and secondary copies of metadata, with the data always still stored on the spindles of primary storage. This should improve both search latency and user logons by 50% according to Microsoft. While this is in Office 365 and Exchange 2016 today, it should be in-built to Exchange 2019.

SSDs are much less expensive today than they were even just a short time ago when Exchange 2016 was released, and will probably continue to drop in price. Having indexes on SSDs should greatly improve client performance when searching for content. Now, I wonder if I can get my entire mailbox onto a RAID 10 array of SSDs?

Microsoft also increased the maximum amount of RAM recommended for Exchange 2016 from 96GB to 192GB. While that is for 2016, you can expect that 2019 will support that much or more. No word yet on whether core count will increase from 24 or not, but we’ll let you know when we find out.

Office 365 already offers dynamically expanding archives for Exchange Online users. It’s a good bet that this capability will be include in Exchange 2019, so users can save even more email soon.

Don’t overlook the client

If you are interested in running Exchange on-premises, you are probably also still using the traditional, MSI-based, client for your users. You don’t have to, since Office ProPlus can absolutely be used with on-premises systems, but on the assumption that you aren’t doing anything “in the cloud,” it was also indicated at Ignite that the next version of Office perpetual, Office 2019, will likely release along similar timeframes, with a preview during the middle of 2018 and an RTM before the end of 2018. As above, while it’s likely that Outlook 2010 will be supported when connecting to Exchange 2019, that won’t be for long, and you really should already be considering an upgrade to your clients if you are still dependent upon Office 2010 for anything.

The wishlist

Here’s what I, and many others, are hoping to see in Exchange 2019.

  1. In-place upgrades for Exchange 2016, and 2013. I know, that’s asking for a lot, but it would simplify things greatly for the SMBs out there.
  2. Easy upgrade of the Exchange org. If you’re running Exchange 2016 or 2013, odds are good adding Exchange 2019 into that org will be full of effort and require transitioning namespaces and more. It would be great if we could start the public preview of 2019 by just adding a server to our existing org, and being able to rip it out or redo it if something doesn’t go according to plan.
  3. Interoperability with Exchange 2010. If we learned one thing with Exchange 2016, it’s that customers are going to have legacy products whether we or Microsoft want them to or not. If going from 2010 to 2019 requires an intermediate upgrade the way 2007 to 2016 did, lots of customers are going to put this off to the very last possible moment, and then do it wrong in their rush to fix things.
  4. A maintenance cycle like Windows 10. Windows 10 is the “last version of Windows” with an update cycle that will keep it current forever; or at least, that is what they tell me. Naming Exchange 2019 for the year of its release (okay, more like cars than birthdays) makes this seem unlikely, and calling it Exchange X would have probably caught a lot of grief from the Twitterverse, but it would be nice to see this as the last Exchange upgrade.
  5. OWA as a full replacement for Outlook. OWA in Exchange 2016 is great, but it’s still not 100% a replacement for Outlook. I’d love to see the OWA in Exchange 2019 be something that would make me prefer it over installing Outlook. That would be a great thing for all the VDI customers out there too.

Wrapping it up

So, there you have it. We know Exchange 2019 is coming out next year. A TAP is already open for it, a Public Preview should hit in the summer, and RTM should be in the fall. We expect it to follow the architecture of Exchange 2016, and include some of the best features of Office 365 in at least some capacity to help on-prem Exchange admins offer more, or less, as appropriate. We predict that there will be little to no legacy interoperability for legacy Outlook or legacy Exchange, and the rest is up for debate. Watch this blog, as we will be sure to include more as we learn of it.

Did you attend Ignite 2017? Catch wind of any rumours or slips of the tongue regarding Exchange 2019? If you did, or heard something elsewhere, leave a comment and let us know what you think. Do you have a wish to see in Exchange 2019? Let us know that too. Thanks!