Can SMBs Build High Availability Exchange Environments on their Budgets?
Recently, Jerry Melnick wrote a great article on eWeek.com on achieving high availability in Microsoft Exchange. He offered five tips to ensure an Exchange environment’s reliability in the simplest manner possible. After reading the article I thought, great advice, but how practical is it for small and medium-sized businesses even with today’s cheaper technology?
Let’s look at some of Jerry’s recommendations.
First, he suggests opting for brand-name hardware. Research shows that brand-name servers command a 15-20% premium over their white label equivalents. A moderately equipped Dell PowerEdge 1950 server carries an MSRP of about $3,000. But you don’t just need one, you need at least three – a database server, a Hub/client access server, and a domain controller, assuming you do not care about any level of redundancy. So, the total premium of brand-name hardware is about $1,500 for the base server alone.
Second, Melnick strongly encourages that storage should be RAID configured, with fiber-channel connections to the Database Server(s). Today’s Storage Area Networks get expensive very quickly, beginning at the low end of $20,000 to over $100k for a basic enterprise entry level unit. Plus, you need to install fiber channel controllers on each server connecting to the SAN as well as fiber switches. Again, not cheap.
Next, he urges for redundancy…everywhere! I agree that redundancy is good and necessary for a high availability Exchange solution, after all who today believes email outages are non impacting? Although putting in extra equipment to sit idly, just in case primary hardware fails or for a software patching gone wrong, is pricey. The cost of redundant routers and switches, firewalls, servers, fiber cards, and more adds up very quickly, especially when considering going with brand name hardware providers.
Finally, Melnick urges companies to adopt full site replication, meaning that all data from the Exchange environment should be synchronized (in near real time) to a completely redundant hardware environment, sitting in another office or data center. That’s another $9k in servers (not to mention networking and storage equipment) sitting idly, just in case. And don’t forget the bandwidth costs involved in keeping the two parallel infrastructures in synch.
I think the eWeek article makes it clear that while building a highly available Exchange environment may be possible for a company with a highly skilled and somewhat large IT team, it is simply not realistic for most SMBs to spend the capital and dedicate the resources needed to manage such a critical application to their business.
Despite this, email is just as, if not more, important to smaller business as it is to the big guys. A report from Eagle Rock Alliance alleges that 40% of businesses that go more than 24 hours without access to their email and data go out of business. Completely out of business, due to an email outage.
If more small businesses knew these stats, they might be a bit more eager to find a better solution than running their stand-alone Exchange server from their break room. This type of information shows that availability, reliability and security – things that businesses are often wary of in a hosted solution – should be, in fact, the biggest drivers for adopting a hosted Exchange solution. Above cost, above ease of management, above anything else. Keeping your email up 99.999% percent of the time is critical to an SMB and it is prohibitively expensive – from hardware, software, staff and productivity perspectives – for smaller companies to achieve on their own.
This is why the software as a service model, and particularly hosted Exchange 2007, is seeing such great growth and will continue in the SMB market over the next couple of years. As a hosting provider with four Tier 4 datacenters, hosting our customer’s environments, we are able to meet and mostly exceed, every recommendation the article outlines for our hardware and infrastructure. We leverage EMC Enterprise Class SAN arrays for our storage infrastructure; we protect against network failures with Cisco 6500 series Catalyst switches along with firewall services and ACE load balancing modules; and we leverage virtualization for those components that are not easily made redundant.
Intermedia has built a highly robust infrastructure for its Hosted Exchange Environments and we spend a lot of time explaining to both current and prospective customers – as well as the media, analysts and other members of the industry – how we leverage this infrastructure to achieve high availability, redundancy, and the best in class service, and why it is so incredibly important within a business.
As a company, we are extremely proud of the time, staff and enterprise-class hardware that we’ve used to build our Exchange infrastructure, because we’re able to provide our customers with something they couldn’t otherwise achieve from a cost and/or technical perspective – a constantly ‘up’ email solution that will never jeopardize their business. Not to mention all the additional add-ons we make available that would normally be very cost prohibitive on an individual customer basis.