Yesterday I received the following comment on my Office 365 Service Upgrade (Wave 14 to Wave 15) Post:
“Turns out You’re another typical Microsoft sycophant. Did you notice the now in excess of 2000 Microsoft Office 365 Forum entries from small businesses like mine that are absolutely B*ll S**t over Microsoft’s treatment of office 365 early adopters using 365 2010 that will have to wait for up to 9 months for their upgrade. It would be nice to see blogs like your’s that I have previously respected actually criticize the handling of this upgrade. From the perspective of small business early adopters we’re being screwed.”
First thing I had to do was look up the word Sycophant I disagree strongly that I apply to the definition of that word, but that is not the reason for this post.
I would like to explain why I feel Microsoft, and other Cloud Providers, are in a tough place with upgrades to their Cloud Services. On the one hand you have customers like the one who posted the comment above. The day that Microsoft is able to upgrade their Office 365 Tenant to Wave 15/vNext it should be done! And any delay is penalizing the customer for being a loyal customer and in the commenters own words “From the perspective of small business early adopters we’re being screwed.” This perceived bias against existing customers being penalized is nothing new, many times loyal long term customers feel slighted by companies trying to lure new customers with price cuts, new features, etc.
On the other hand if Microsoft did upgrade every one of their existing clients to Wave 15/vNext of Office 365, they would be receiving a similar backlash from clients that are comfortable on the existing version and want to wait before being upgraded. These type of clients want to control the change within their environment, even when it extends to the cloud. These clients would rather rely on other customers upgrading first and learning from those who go first how the transition goes and then when ready have their Tenant upgraded.
I think a move to the cloud needs to be examined by companies and really the company needs to ask itself, “With a move to the cloud am I willing to give up some of the upgrade control I currently have with my services”. While this is probably not the most important question regarding a move to the cloud, I think it is one that is not really thought of until a situation like the one we are talking about now arises, upgrade of Office 365 from Wave 14 (2010 versions) to Wave 15 (2013 versions) services. This is an important questions because with a move to the cloud, the company give up some control on when, and even if, you will go through an upgrade of the service. The company now relies on the Service Provider, Microsoft in this case, to handle the upgrade and the cadence of the upgrades. This needs to be fully understood and accepted by a company moving to the Cloud. When a company runs the services on-premises, they have complete control as to when and even if they will upgrade to the latest version. An on-premises company might decide that new feature X is so compelling that they must upgrade immediately, or they might find no compelling reason to upgrade at all and totally skip a version.
So back to this upgrade cycle of Office 365 from Wave 14 to Wave 15. I am sure if you asked Microsoft, they would have loved to been able to send out an email on February 27th asking each customer if they wanted to upgrade now, in four months or in nine months. Then based on the responses upgrade the customers according to their wishes. This of course what not possible, and without going into details I know and probably more that I don’t know, it is just not possible. Microsoft has numerous datacenters and so many servers it is probably hard to get an accurate count on any given day due to all the provisioning and de-provisioning going on daily. Ensuring that the services are up and running and eliminating downtime and service interruption is at the core of running Office 365. They have a Service Level Agreement (SLA) of 99.9% uptime that is financially backed. So Microsoft has to ensure the upgrade process to get customers to the latest version does not impact the service they are providing.
Side note on this, I find it humorous that companies that move to the Cloud immediately believe that running Exchange is so easy that an outage should never happen. I have personally seen companies that I know struggled keeping a single Exchange Server running on-premises for 3 days at a time being the loudest complainers when an hour outage is suffered by Exchange Online. Microsoft does strive for 100% uptime, and even with planned maintenance does come closer to this that most any other individual company could come close to running the services on-premises. So guess what I am saying, is yes it is big news when Office 365 suffers an outage or even a service degradation, but cut some slack and think back to when your company was running the services on-premises and remember the challenges keeping everything running 100% of the time!
To upgrade a customer in a multi-tenant environment requires that not just the customer’s primary servers that host them are ready for upgrade, but that the secondary servers within the customer’s datacenter as well as the secondary servers in the alternate datacenter are also ready for upgrade. Remember, Office 365 enables a customer full datacenter redundancy as well as Disaster Recovery to an alternate datacenter. Again prepping and ensuring the servers are ready for upgrade are done with the #1 goal of no service outage or degradation. Someone might ask, why not just buildout a completely new environment that could host all of the existing customers and move everyone over to it? Well that is great in theory, but that costs money, and one reason most companies go to the Cloud and specifically Office 365 is that the price is right. For Microsoft to need to double the amount of servers, storage, power, cooling, etc what do you think that would do to the price of the service?
I guess getting back to my point about upgrades, is as a company with services in the Cloud you give up some of the control you had when running the services on-premises. You need to understand this. If you can’t accept this, maybe the Cloud is not a good fit for you, or you are not a good fit for the cloud, whichever way you look at it. I am sure this explanation with lead my commenter above to still think I am a sycophant, and I will still disagree, but many things go into the process of providing a robust, highly available and upgradable Cloud, and I am just trying to get Cloud customers to understand and realize that there is a bigger picture than just your Tenant in the Cloud!