A few weeks ago, I wrote about the results of a poll on the Exchange & Outlook page about moving messaging to the cloud. The results of that poll showed that the Exchange Server admins in our audience are fairly heavily biased against this model, whether Microsoft or anyone else is running it. As a result of that post, I received some reader feedback explaining in more detail some of the problems with moving to the cloud. In addition to the typical concerns—security, control of data, fear of outsourcing—it was also mentioned that the cloud model is seen as still too new and doesn't yet have all the features of legacy on-premises systems.
We'll forget for the moment that some hosting providers have been offering hosted email for 10 years or more—which is nearly an eternity in the tech world—because the point is still valid as long as that's the perception of the IT pros in the field who are being asked to deploy or work with such systems. And furthermore, when the hosting providers still regularly have widely reported outages or data leaks, it certainly lends weight to any argument that says the technology isn't quite ready yet. In fact, on Sunday, Microsoft Business Productivity Online Standard Suite (BPOS) experienced yet another outage, as reported in its Twitter feed from msonline:
E-mail service issues for some in N. America. Team is working on it. See http://tinyurl.com/23pb5xz for latest. Sorry for inconvenience.
Nonetheless, some businesses will still find using a hosted service for their messaging systems the way to go. I received a response from a reader on this point, as well, and in fact he described in detail how his organization compared the numbers between upgrading to Exchange 2010 or moving to BPOS and found the cloud to offer both cost savings and additional features that their current system didn't have. The reader asked to remain anonymous, but he has let me share his message, which is copied below. If you are considering a hosted solution for Exchange, this can serve as a good outline of some of the things you'll need to factor in to your decision. I hope you find it useful.
Meanwhile, here's a little something else to keep in mind if you're still on the fence about cloud computing—or even if you don't like it but are being asked by company execs to figure out how to implement it. We'll be putting on a cloud computing conference in the spring with a full slate of sessions by the experts that can help you understand the benefits as well as see how the IT pro's job shifts but doesn't end with this kind of move. Stay tuned to this space for more info on that show as it becomes available in coming weeks.
A Reader's Response about Moving to Hosted Exchange
We are currently planning a move to the MS BPOS system. The company I work for is a large electrical contractor and has offices across the west coast. We are still using a centralized Exchange 2003 system and have seriously considered the costs of upgrading to Exchange 2010 vs. moving to Google or BPOS.
I think the first thing a company should do is consider what the objective of the upgrade is. For us, it's to provide larger mailboxes for users to store data, better performance (slow opening large attachments over remote links), move off of old software/hardware, and disaster recovery.
A larger mailbox is really important to us since we deal with a lot of large email attachments. We currently use an email archive solution that stubs old messages off to secondary storage. It works well, but it's slow (stubbing job runs like a brick and mailbox backup takes 12-18 hours—ouch), takes time from the users to go through their mailboxes and mark items to archive, and requires a plug-in to be installed in Outlook that is finicky.
What we have found is that the initial cash outlay for the upgrade in our datacenter is $170,000 (additional server hardware, SAN disk, and licensing) with an additional $100k for updates to our co-lo facility.
Moving to BPOS is roughly 7k a month for the number of users we have and features we want. This enables up to a 25GB mailbox, adds chat capabilities and two-way video conferencing through OCS, up to 10-year archival, built-in DR, and SharePoint for file sharing. Since access to email would be solely dependent on our offices' Internet connection, I am required to add an Internet circuit in each of our branch offices (currently their Internet traffic goes across our MPLS and out from our DC), and a proxy server and firewall at each location. Which I think is OK as it will reduce some traffic on the WAN circuits.
So to recap the numbers:
$270,000 hardware and licensing
4 months labor for the conversion (this company considers IT labor to be free since we are working here anyway . . . blah).
$80,000 annual costs for hardware support and co-lo fees.
Three year cost: $510,000
$38,000 from proxies and firewalls
$15,000 2 months conversion time with consultant help. Mainly needed help to recover the stubbed email and copy up to the users' new mailbox.
$3,600 per month for business class DSL/cable Internet circuits (3 x 100 x 12)
Three year cost: $308,600
So just looking at the $$ it's fairly easy to see that it is cheaper and provides more features to the user than an Exchange 2010 upgrade. It also simplifies our environment considerably, saves money on hardware & software support costs, and a lot of worry.
Downside is that public folders are not supported in BPOS. We plan on replacing them with SharePoint sites, but currently you cannot mail-enable a BPOS SharePoint site. Hopefully by the time we implement, this feature will be enabled.
I think of it as going back to using POP/SMTP for email but with all the Exchange contacts, calendaring, and OCS features added in. Back when we first started using email (97/98) you really never worried about the remote server going down (assuming you used one of the major providers); more often than not it was the Internet connectivity that caused an issue with getting the mail.