All "clouds" are not the same

With the recent disruption of some high profile websites on Amazon's EC2 platform (Reddit, Foursquare, and Moby to name a few), some naysayers are pointing to this as a weakness of cloud based services in general, rather than what it really is, simply a single provider having issues with specific services.

While Amazon locates servers across multiple "availability zones", the fact remains that a single rogue "network issue" was able to bring many associated zones offline simultaneously.

In contrast, we (Greenview) have distributed our servers in such a way that we are not dependent upon any one data center, or even any one provider. Even the famous 2008 data center explosion at "The Planet", where we have many servers, had no noticeable affect on our service. Back in 2003 when our home office and much of the East Coast lost electricity for several days, our service continued uninterrupted. The only issue was that we couldn't update our antispam databases and therefore let a bit more spam through. Our customers without electricity didn't even lose email because we automatically spooled it for them. During these and other data center failures, our service has been unaffected due to our multiple, geographically disparate data centers spanning many different providers not just across several states, but across several countries.

We anticipated and learned that despite advertised claims, even the most reliable and redundant servers inside the most "State of the Art" data centers, can still fail. We kept these lessons in mind when designing and implementing our new Zimbra hosting packages. Our Zimbra clusters have multiple servers per data center, replicated in real time across to a duplicate site with another identical set of clustered servers. We could launch new instances in either data center, or change data centers completely, within minutes if absolutely necessary. This may sound similar to Amazon's claims that "Availability Zones are distinct locations that are engineered to be insulated from failures in other Availability Zones", but we are using two completely independent providers. And that's just part of the package.

Additionally, every hosting package comes with our 30 day Email Archiving service included for free. This adds not just additional copies of your messages, but adds a completely independent infrastructure, to aid in our disaster recovery plans. Our email hosting customers have the option, for free, to archive the last 30 days of their email into a completely separate service spanning three data centers each with their own internal replication amongst local servers.

What does all this mean for you? More redundancy means higher uptime, and fewer worries. In fact, our services have had a 100% uptime to date since our launch in 2002. Additionally, economies of scale allow you to have these enterprise level features at a rock bottom price, a fraction of what it would cost otherwise. Add to this robust infrastructure our 24/7/365 support team, entirely based from our Ann Arbor Michigan headquarters, and you have a service that you can count on. While no service can claim to be "bullet proof", using multiple providers in geographically disparate data centers gives us a distinct advantage over in-house solutions, single server hosted solutions, or even multi-server solutions that rely on a single provider.

While none of our servers are housed in Amazon's data centers, our thoughts are with their team and their customers as this issue sorts itself out.

For more information on our services, our infrastructure, or just to test our 24/7/365 "No Voicemail" support service, please feel free to give us a call any time.