I can't tell you the number of Windows Clusters I've implemented over that past fourteen years because of patch Tuesday, but let's just say it was more than I could keep track of. If these words ring true for you, you'll be very excited to hear that EMC is releasing a massive Solution white paper that discusses how to bring your systems, not just up to local availability, but to Corporate Availability. We are bringing you abilities to not just patch whenever you want, but to do whatever you want, whenever you want — at all levels of the Storage/Compute/Network stack. We are bringing you products and techniques that protect your data and your systems better and more accessibly than ever before.
But what the heck is Corporate Availability? It's the level of application availability that business units have learned to expect and demand. It means that the systems they use to generate revenues, to manufacture products, to maintain contact with their customers will be available -- constantly available. Corporate Availability means that local High Availability techniques are no longer "good enough". It means that patching needs to occur without a testing cycle -- it means that the systems that have been patched need to be able to go backwards in time to "undo" a patch that my have negatively impacted the system. Corporate Availability means that, no matter what is happening in Server Administration, Network Management, or Storage Management, every application owner can focus on running their applications. There's no more tolerance for scheduling outages. There's no more "agreeing to wait for them". Application owners need to have access to their application infrastructure constantly and consistently. That's what Corporate Availability demands.
There once was a time when applications had "maintenance windows". I remember when I was a Technical Account Manager (TAM) at Microsoft back in the last decade, the customers I supported were large, international financial services firms. Their IT shops, just like today, need to support hundreds of SQL Server instances, dozens of Exchange 2003 servers, and hundreds of Windows File Servers. Every month, they would wait for Patch Tuesday. At about 10:00am PST, the email blast would come from Bellevue or Redmond announcing to the organization, which groups would need to begin their systems tests based on the latest hotfixes for the "following products are affected" list… and the fire drill would begin. Who would be affected, how long would the systems be down, who would work late, what impact would these updates have on related systems. How would we role back?
Many of the issues related to "Patch Tuesday" have been satisfied by Clustering techniques and by snapshotting technologies — both VM's as well as datasets. Hardware VSS has made going backwards in time a commonplace concept. But what about the rest of the stack? Can application availability be totally satisfied by snapshots? We all know the answer is "NO".
Networking and the "connectedness" of our user base is so complex, and so entirely required for everything we do, that we need to provide layered techniques to maintain access to our valued datasets. These datasets are, of course, the reason that IT organizations and IT infrastructure exists. Each data element, whether it's a customer order for a pair of black and white checkered Vanns skateboarding sneakers or an email from your HR department outlining the new holiday schedule for 2015, those data elements are the key to an entire system's existence. Allowing the consumer, the end-user, to access those data elements is the stuff this new EMC Solution is all about.
In the Solution, we gathered your favorite applications -- ok, so "favorite" might not be the right adjective… -- we gathered your most popular applications -- Sharepoint, Exchange, and SQL Server. We implemented them in the way that you would most likely chose. We even put the SQL Servers into Failover Clusters, not AlwaysOn Availability Groups. And we demonstrate how to maintain application availability across THREE sites. EMC's architecture does not REQUIRE three sites, but it allows you the opportunity to explore what EXTREME availability can look like. And it's pretty amazing how simple the whole thing really is. It's not simple because it's 2014; it's simple because it's EMC. The folks you've trusted since 1990 when we brought you Symmetrix, the world's first storage supercomputer.
Please take some time to glance through the new white paper -- it'll be launched very soon -- it'll show you the exact steps to take to create an availability infrastructure unlike anything you've seen before. It takes advantage of products that you have come to know and trust -- components such as EMC VNX, EMC RecoverPoint, and EMC VPLEX. It uses new code versions that allow them to all know about each other and do things that you've wish they could do. Things such as replicating from a stretched Windows Failover Cluster (located in two synchronously-connected data centers) to a third data center that's asynchronously connected (three states away) AND automatically fail to it if the network access links get a visit from the Grim Backhoe.
This blog will be updated when the paper is published, but for now, please post your comments below. Please let me know if you are as excited about this paper as I am!