How Can Different Upgrade Methods Help or Hurt Your Backup Solution’s TCO?
In the Backup and Recovery world, there’s no shortage of acronyms and buzzwords used to refer to spending and savings. CapEx, OpEx, TCO, ROI, etc…every vendor will use these to make their solutions seem more attractive from a cost perspective. However, when you dig a little deeper into how a product is installed, maintained and more importantly, upgraded, you can start to see where hidden costs are, and maybe that attractive “Total Cost of Ownership” you discussed back when you first purchased the product, isn’t that pretty anymore.
Upgrading and expanding your backup software’s footprint to meet your ever growing data center comes at a cost. This is an expense that is often overlooked when the initial purchase takes place. A product may seem less expensive than other competitors up front, but as the old saying goes, “the Devil’s in the details”.
I want to focus on two different approaches when it comes to scaling you backup and recovery software. There’s the “monolithic” (scale up) upgrade, and there’s the scale out method.
The two-tiered approach (backup server and nodes it backs up) may seem attractive and simple at a glance, but scalability becomes complex and costly as a data center grows. The backup server in this case is the sole owner of the proprietary database which is the brains of the entire backup operation. When your environment outgrows what one of those servers can cover, what do you do? You simply buy another backup server and assign it its own nodes to back up. Now you have two backup servers, two separate databases and separate groups of nodes covered by each server. Imagine where the simplicity goes when you have to do this a few times a year, each year? Not only does it cost from a software license perspective, but because these servers are incommunicado and their nodes lack collaborative integration, you have all these separate backup environments in your data center. Without a single pane of glass and centralized database to manage these, you need more administrative help to make sure your backup environment is running as it should. Roughly one fifth of the Backup and Recovery market run their day to day operations with this type of architecture. That’s a lot of dollars and administrative overhead!
There’s a better way.
The three tiered backup architecture does not back you into a corner with the above problems I just discussed.
Here’s how the three-tiered solution works. The brain of the entire backup environment is a main server or “master” server. This system is responsible for managing the scalable database and catalog for all backup policies in the protected domain. It manages the second tier of what are called “media’ servers, who are actually the workhorses that facilitate the backups of all client servers. The master can actually act as a media server itself, but in larger environments, they tend to leave that to the media servers under it. This provides a huge advantage when you talk about scaling your solution to fit your growing data center. You can continue to add media servers in this type of architecture, and still manage everything with a single database and master, as well as from a single pane of glass. No need for more databases and additional “master” servers. “Scaling out” allows this to be easy and painless, and you don’t have to keep adding headcount to manage the environment. This architecture also allows the 2nd tier (media servers) to work together with load balancing, making your backups run most efficiently, and complete in less time.
Some will ask, “Isn’t buying more backup servers in a two-tiered environment the same cost as buying more media servers in the three-tiered option?” The answer is simply no. The main backup servers manage databases and catalogs. Each one needs its own. Media servers don’t. They are easier and less expensive to add, they don’t create separate and independent database sprawl, and they are all easily and centrally managed. This cannot be done with the traditional two tiered way, which significantly increases your OpEx and total cost of ownership, as well as diminishes your ROI.
Make sure you really know what you are going to spend on backup. Not just what the original P.O. says, but each time your data center dictates it’s time to “expand”.