Thoughtsofanidlemind's Blog

“Be afraid”, be very afraid”… is the opening of a blog post by Scott Cameron, an account strategist for Microsoft, titled “Office 365: The End of Exchange?” It’s an odd post that combines a fair element of truth with some fairly dubious predictions. This is a topic that I am asked about often, so it’s worth a comment or three. I do this from the perspective of using Office 365 for my business and working with many companies who are quite happy with on-premises Exchange.

I agree with Scott’s assertion that messaging engineers and administrators need to build their cloud skills (or rather, their experience and knowledge of the technologies associated with cloud-based software) and that “you’re not building your skills you’re falling behind.”  This is good advice given the state of play within the industry and the rapidly growing acceptance of cloud platforms as an option for utility IT services like email. It’s also true that IT people who don’t continually build their skills do fall behind. We have seen this aspect of the industry play out many times in the past as IT has transitioned from mainframes to minicomputers to client-server and now cloud. About the only set of people who have managed to stay constant are those who write COBOL for mainframe applications, and there are precious few of those.

I could be pedantic and say that email timeline outlined in the post is flawed (in the early 90s, many people had email; for instance, DEC’s ALL-IN-1 Office System had more than 5 million corporate users; Gmail launched in 2004, not 2007; Office 365 launched in June 2011 not late 2011…), but that is not the heart of the argument.

I don’t agree with his view that on-premises email servers are in imminent danger of becoming extinct, nor that Office 365 is a universal panacea for all of the ailments from which on-premises Exchange might suffer. I have written about this topic before, stating a case why I don’t think on-premises Exchange is dead (and won’t die any time soon) and also pointing out that the success of Office 365 is still based on relatively small numbers when compared to the overall installed base.

Even Microsoft VP Perry Clarke, head of Exchange development “agreed that somewhere in the region of 40 percent of the total installed base will likely continue using on-premises software for at least the next several years” when I interviewed him last December. It’s true that Office 365 is continually growing and that Exchange Online has expanded by 600% since 2012 to now require a pool of 100,000 servers, but even so, the word coming through at the Microsoft Exchange Conference was that the next major version of on-premises Exchange is on track for delivery towards the end of 2015. Given the normal support lifetime of any Microsoft server application, we have evidence that on-premises Exchange will therefore be around for quite a while yet.

Although on-premises Exchange will persist, the pool of those running on-premises servers will decline over time, simply because it makes terrific sense for many companies to purchase email as a service. Why, for instance, would any start-up company consider installing their own Exchange server today? I also think it true that most small to medium companies (anything under 1,000 seats) will find it difficult to justify the expense of running their own email servers. It makes more economic sense for them to devote their time and energy towards building their business and creating added value rather than wondering how to deal with the next cumulative update for Exchange.

For this reason I agree that those running “vanilla” on-premises servers today need to increase their skills so that they can deal with hybrid deployments or a migration to the cloud. And after that migration is complete, the administrators need to understand what role they have to play in the new world. That probably means that new skills have to be acquired because many of the old skills are not needed when servers are managed within Office 365. It’s good to have a strong job continuation program…

But there are many companies that cloud services for whom cloud services do not make sense. At least, not in today’s world. Things might change – and will probably change – over time, which is good reason to keep an eye on what’s happening, but for the moment some difficulties and challenges exist that allow a fair argument to be made against moving to the cloud. The list of excuses for not embracing the cloud described in the blog is amusing and valid in some cases, but real-life and relevant reasons do exist when embracing the cloud does not make sense.

For example, companies who are unhappy about the PRISM revelations and the notion of losing control of their data are unlikely to want to embrace the cloud today. Microsoft can argue as much as it likes that they do not allow unfettered government access to data, but the point is that people believe that this happens. The FUD principle has always been a useful one to deploy in making a case for one technical choice over another.

Networking is another issue. Some companies – even in highly developed nations such as the United States – operate in areas where copious amounts of high-quality bandwidth cannot be secured. Some of these companies find it difficult to network operations today; the prospect of moving everything to the cloud leaves them cold. I am reminded of this at every conference I attend by people who are in this position (normally after I’ve just told an audience why cloud services are good).

Evergreen software is touted as a great advantage of the cloud. And it is, but only for technologists. Normal users (and those who have to support those users) are often unhappy when user interfaces change. Microsoft promises to make many changes in Outlook Web App over the next year to introduce new features like “Clutter” (one that I am looking forward to), but the prospect of seeing new machine-learning technology being suddenly revealed to end users is a cultural change with which some are still uncomfortable.

Support remains the Achilles Heel of the cloud. First level support is often script-driven and tiresome. It is a necessary evil that has to be endured before you can be escalated to someone who might understand what is happening. Even then, because of the scale of the cloud and the absolute and understandable need to restrict change to the absolute minimum (to avoid adverse impact on other tenants), even second level support cannot do much in a practical sense except gather evidence to support a case for further action. In short, the Office 365 support experience can be very slooooow…

My point is that every computing platform ever developed has both advantages and disadvantages. The trick is to maximize one and minimize the other as you make technology choices for your business. The table below provides some areas for comparison. As every company is different, you need to put this data into context with business needs, operational and IT requirements, economic situation, and available skills to come up with an answer for a particular company.

Office 365/Exchange Online	On-premises Exchange
Security (many certifications attest to this)	Security and Privacy (tailored for the company)
Predictability – what the cloud vendor says you get, you get	Flexibility – you can tailor the software with third-party or bespoke code to meet business requirements
Evergreen software deployed as the cloud vendor dictates	Control over when and how client software is deployed
Slow, script-driven support	Responsive, user-sensitive support
Fixed monthly cost per mailbox (but don’t forget the extras, like ADFS for hybrid deployments)	Tighter control over cost (possible, not always practised)
Data secured through native data protection	Data secured through native data protection and traditional backups
Financially guaranteed SLA (as measured by Microsoft)	SLA remains in the control of the IT department (possibly worse than achieved by Office 365)

As Scott concludes in his post “Are you ready for the cloud?  Time’s up.  It’s here.” Yes it is, but it’s been here for a while and some of the issues remain unsolved. That’s why on-premises Exchange and its strong hybrid capabilities remain a key player for Microsoft in this market. Clouds float. The right one will come along – eventually.

The other thing to remember is that a post written by someone who works for a supplier is always prone to bias. The author might not be intentionally biased and attempts to write without bias, but the fact that their pay check is linked in some way, shape, or form to success in a particular area creates huge potential for bias. And given that Microsoft is going through a “cloud first, mobile first” upheaval at present, it’s reasonable to expect that the corporate Kool-Aid is being consumed in a major way by many employees.

It’s always been the way… well before cloud services ever appeared.

Follow Tony @12Knocksinna