Today's post is the first of two guest posts this week by Julia White, Sr. Director of Microsoft Exchange Product Management.
For those of you who don’t know me, I run the Exchange product management team at Microsoft. I often post to the Unified Communications Group Team Blog; but my colleague, Tom Rizzo, suggested I guest blog here. So today and tomorrow, I’ll be posting my thoughts about email, on-premises and in the cloud.
As part of my role, I talk regularly with customers about their email solution. Nearly all of the conversations I’ve had in the past year include a discussion about cloud-based email and how IT departments should think about their next upgrade. The discussions appropriately center around three key areas: control, capabilities and customer support & service. Within the area of control is where discussion of uptime and Service Level Agreement (SLA) comes up.
Get the Truth on Uptime and SLA
Google has attempted to take jabs at Exchange uptime as compared to Gmail and as such, I wanted to take this issue head on. Email is considered a mission critical application at most organizations, so uptime is certainly important and worth understanding. You may hear Google tout that Gmail’s uptime is much better than on-premises Exchange. Through their claims, it is clear that Google is clueless about on-premises email and are insulting to the dedicated IT Pros who run these systems. But, beyond that, I want to make sure you have an apples to apples comparison of uptime.
The majority of Exchange customers have either recently completed their upgrade or are in the process of upgrading to Exchange 2010. Utilizing the latest tools available in Exchange 2010, our customers are seeing up to 99.999% uptime on premises. On the services side, Exchange Online customers get a financially backed, 99.9% uptime guarantee. If you are considering Exchange for your next e-mail system, these are the only two versions that are relevant to you. Make Google talk straight to you – and stick to what they know, which clearly isn’t on premises email.
Customers Need Feature Predictability...with More Than 1 Week Notice!
Getting to aspects of control that are beyond uptime and SLA, a key concern I hear from customers is the predictability of a cloud email service. Over the past 10+ years, Exchange has provided a steady, predictable set of innovations, so it’s no surprise customers have learned to rely on this. We know the cadence of our product release cycles is important to large enterprises and small businesses, because it provides a predictable series of events that our customers can plan for and depend upon. For enterprise products, we generally disclose upcoming features in the next version of the product at least 12 months in advance. We hold a public beta that allows people to test an early version of the product and provide feedback. At the end of this process, we release a fully tested, finished product. Within the next year, we make available a Service Pack that is also made available as a public beta, with minor enhancements and a rollup of bug fixes. We maintain the same disciplined cadence whether we’re releasing an on-premises or online product.
Google takes a different approach, which serves Google well, but not customers. Recently, Google announced some modifications to their release cycle with two tracks: those who get access to features immediately and those who get at least a week’s notice. While it is clear Google is trying to show they are beginning to understand the enterprise, it is also clear they have a long way to go. Not all features will adhere to the new release cycle and not all features will be in beta or even final versions, nor is there a timeline that shows when those features will be pushed out, without a choice. This isn’t the first time customers have experienced ambiguity from Google. Gmail was in beta for four years, Google Wave came and went, Google Docs was available offline with Gears and then offline access was removed. Can you trust Google with your business, and will your employees trust such a haphazard feature release strategy? Google’s release cadence is convenient for Google, painful for consumers and downright disruptive for businesses.
As we release online products, we follow the same predictable process we use when releasing major updates, but with the added benefit of minor software updates, delivered on an ongoing basis. Plus, current Exchange Online customers will have the option to schedule an upgrade to Office 365 for up to one year out. That is the type of control we know our customers need.
Does IT have All Tools Required to Effectively Manage Cloud-based Emails?
Another area of control I hear a lot about is e-mail administration and this comes down to tools. With the upcoming release of Office 365, our customers have the same features available in Exchange Online as they do in Exchange Server:
For those of you who want to read more about the features listed above, visit the management tools page on the Exchange website. With these tools, you have the same rich features in the web interface, and you can use or modify them just like the administration console in Exchange Server.
Check back tomorrow, when I will discuss the flexibility and familiarity you have with Exchange.
Julia WhiteSenior Director, Exchange Product Management
"Utilizing the latest tools available in Exchange 2010, our customers are seeing up to 99.999% uptime on premises."
The only recent situation I've experienced with a multi-day total email outage for over 1,000 users was with Exchange. In that situation, the Exchange computers (yes, multiple failovers) needed to be reinstalled and the database backups restored from tape. This was a government installation which surely blew hundreds of thousands of dollars in taxpayer money in lost productivity thanks to on-premise Exchange. Granted, this was the 2007 version and not the supposedly more stable 2010 version with latest tools. As this was a four day continuous outage, it put us at 98%.
In the small business I work in today, I inherited a similarly flaky Exchange server. In this situation, I immediately implemented granular backups and failover. Unfortunately, our power generator is only powerful enough for critical operations with 8-core power-hungry computers not considered critical. Several times in '09 and '10, the power would go out long enough to drain the giant 4U UPS. Long story short, even though we had plenty of phone battery and backup, we couldn't send or receive email during these times.
Worse yet was the configuration of this Exchange server (done by a previous consultant) was unstable in that restarting the machine required manually restarting the Exchange server. I fixed this after a weekend of outage due to a reboot but things like this should not happen. I figure I am not alone in Exchange not being "walk-away" stable.
Although switching to gmail was disruptive from an interface and training point of view, this was more than made up for by cost savings on labor, on-call emergency consultants, and lost productivity due to downtime.
Perhaps one can get 99.999% with tens of thousands invested into on-premise infrastructure, battery backups, and electric bills for multiple servers. As far as I can tell, this won't happen for $50 / user and certainly won't happen with the IT departments in most small to medium businesses or government agencies. We've all heard the expression once bitten, twice shy... well, Exchange has bitten me twice so you may understand why I am skeptical that the latest tools and Exchange 2010 solve any of the basic problems that I have experienced.
I'll tell you what's "disruptive": that "recent tweets" thing on the right-hand side of this article. It moves way too much and is way too distracting for me to be able to concentrate on the article itself. You lost a potential reader.
I agree with Seth. Please remove of any kind of annoying animated web sidebar thing. This isn't 1997, there is no longer a question as to what web users will love or hate.
The definition of irony: the King of Control - when it comes to locking customers in with expensive licensing and software that only integrates with their own stuff, driving mandatory 3 year upgrades/version updates/weekly security patches, etc (should I go on here?) - talks about trust and giving their customers "the type of control they need."
"Mac attack," I don't think the patch auditing is a big issue anymore with Microsoft. I understand the Intune program allows consultants to manage small installations via cloud infrastructure.
The only major glitch I've seen from updates is not being able to delete some large files. I had this happen right before the storage space update was rolled out.
I believe the two track Google system was largely because the docs list changed so radically. One day, Google Docs looked like Gmail. The next day, it looked completely different.
I prefer the way Smartsheet rolls out changes where big info bubbles pop out of updated interface controls. I wonder if Office 365 will have an option like this.
Microsoft and Stratus have been partners for a long time. Not long ago, Stratus guaranteed 100% uptime for Windows running on our servers, backed by $50K. That's how much we respect your products for mission-critical applications. If you could, please shed some light on the basis for 99.999% uptime on premise, and why the SLA disparity on the services side. I've been reading that IT shops can actually achieve better uptime from cloud services than they get in their own data centers (pretty scary!). Just want to hear your take on it. Thanks.
I’m very sorry you were in the unfortunate situation of troubleshooting a configuration that didn’t work for your organization. A four day outage is not acceptable under any circumstance and we hope you were able to find the resources to help the organization prevent this type of outage in the future. I recognize that not every organization invests appropriate resources to plan and maintain their Exchange environment in a highly available manner. For organizations in this situation, Exchange Online is a great option where you are assured of a highly available Exchange environment, while maintaining the same IT Pro and user experience of your Exchange Server. Also, we have invested in Exchange 2010 high availability and disaster recovery capabilities to lower the investment required to maintain high uptime. For others looking for best practices on running a highly available Exchange Server, here are some resources:
-For Exchange Server 2010 customers, you can find best practices guidance here http://bit.ly/hcXpCf
-If you have Exchange 2007 deployed, you can take a look at these best practices http://bit.ly/dTDVNw
-If you are on Exchange 2003 or earlier and prefer to stay on-premises, we recommend upgrading to Exchange 2010. With Exchange 2010, we’ve built high-availability and disaster recovery features right into the product and the payback period of the upgrade is less than 6 months (http://bit.ly/706SHd)
-If you prefer to move to the cloud, you can get Exchange Online for $5/user/month. Microsoft has made significant investments in infrastructure, people and process to ensure that our datacenters provide the highest levels of availability, and we financially guarantee 99.9% uptime
Ian, I’d like to personally invite to you back to try Exchange Online and/or Office 365, we think you’ll be pleased with the service experience and capabilities.
@Ken, thank you for your partnership. We hope that the best practices in Julia's comment above will help you maximize uptime for your Exchange server as well.
I did sign up for the Office 365 beta although I haven't heard back yet.
I'm also now using intune to manage some client laptops. I hope that this will someday integrate with the other tools. Hopefully also without having to use silverlight which is a bit slow performance wise.
At this point though I would need the office feature set to exceed Google. Things like priority inbox, smart labels, and secure-data-connector will need an office 365 equivalent.
Oh, and the online ms would either have to integrate with smartsheet (the way outlook did prior to 2007) or make an online version of project that had all the formulas smartsheet does. Until then, the best I can do is augment Google Apps. I come from the perspective of already having gone through the scary migration week and now having users used to publishing measurements on smartsheet or viewing ERP summary reports in google spreadsheets.
Is there a way to turn services off and on in office 365 similar to Google Apps administrator dashboard? For example, to turn email off? I guess I should ask the help forums.
Lets be perfectly honest here - exchange 2010 is rock solid and has great HA and DR options. In order to experience a 4 day outage, it must have been designed, implemented and supported by absolute muppets.
Microsoft do plenty of things badly - but Exchange 2010 is not one of them.
While 2007's HA is poor in comparison - the DR processes are still solid - so a 4 day outage indicates there is a severe lack of basic knowledge, not a poor product.
@Ben: Another of Microsoft's strong offerings is Office 365, which includes Exchange Online. Microsoft’s practices in calculating service availability and addressing availability concerns show that the company values its Office 365 customers. Microsoft’s accountable, financial remedies show it is committed to providing strong service availability.(http://bit.ly/swjQpm)