OTC builds a virtual data center that spans power grids
OTC is buildout our a "virtual data center" by knitting together our top tier data centers, located in separate power grids, together with fiber.
Using OTC's "virtual data center" provides improved connection against large scale blackouts. Production servers can exist at two separate data centers, 60 miles apart, each on separate power grids.
Using available networking technology the load can be spread across the production servers at the 2 locations.
"Virtual data centers" offer two distinct benefits:
a) Grow incrementally: Ever been in a data center that's out of space, power or cooling? What do you do next? With a virtual data center architecture your not limited by the capacity in your current location.
b) Uptime: Your equipment physically exists in multiple geographies but are connected as if in the same data center so you're better protected against wide scale power disruption or weather anomoloies.
Every organization with data needs some type of DR strategy
Just about every organization is creating data at a faster and faster pace. At the same time, the value of that data is growing. How many organizations out there could operate (or even survive) if they lost ALL their data, or just couldn't access it for days or weeks.
Data is one of the most valuable assets an organization has, and there's no insurance plan that will cover recollecting it. It's as valuable as a money and belongs in a facility like a bank, i.e. a bank for data.
There are 3rd party data center providers who are much like a bank for your data. You don't keep your cash in your closet, so why would you put your data assets there.
A summary of challenges to high availability in your database
High availability implies little to no downtime. Downtime results from security breaches, lack of scalability or failure (hardware or software). To achieve high availability you must manage all three characteristics for each layer in the Internet Delivery Stack.
One area of real challenge to this is the Database Layer. This article explains a bit why...
Traditional attempts at High Availability involve replicating information to secondary equipment or locations to be used in the case of failure to the production environment.
This replication though represents significant challenges when it comes to the the database layer of the stack. Because databases often change rapidly and require immediate response, it's difficult (i.e. expensive) to replicate the data to secondary hardware or to another physical buliding.
Enterprise SQL servers (such as from Oracle and Microsoft) provide tools to accomplish replication. This expensive and complex software can allow a collection of physical database servers, even in separate locations, to appear as one large database to the application.
If you're looking for a high-availablity environment make sure you invest properly at each layer in the stack. A highly scalable network doesn't help high-availibility if your database server is prone to a single hardware failure.
