This article can also be found in the Premium Editorial Download "Storage magazine: Using two midrange backup apps at once."
Download it now to read this article plus other related content.
Make sure you aren't using your priciest options to accommodate changing data.
For example, company ABC is exactly like your company. It sells things and runs business systems. Employees do payroll, send
| invoices, collect money and do online transactions. There's order processing and ERP, and packages and customers are tracked. Documents are written, code is produced and advertisements are crafted. The firm has a Web site, a forum, a service portal and a knowledge system. It runs collaboration tools and telephones. Information of every description is kept online and ediscovery requests are received. It also has lawyers, accountants and auditors, just like your firm.
The organization's business systems run on top of Oracle, just like yours. It runs mainframes, Sun boxes, Hewlett-Packard, IBM or whatever, just like your company. That data is stuffed into EMC or Hitachi Data Systems just like yours. There are distributed servers, as well as NAS systems from NetApp; and there are small SANs with Windows Server, SharePoint and Exchange, just like at your company. Employees deal with workstations and WANs, remote offices, moron bosses and idiot users just like at your firm.
And for each of its apps, company ABC has an infrastructure that tends to be the same stuff it used when the data was created, as well as a "never-going-to-change-and-let's-hope-we-never-access-it" stage. Just like you have. But what people forget to do is pay attention when data moves from stage one to stage two, and then again to stage three. People leave data that's no longer "dynamic" on their most expensive infrastructure, and apply their most stringent (expensive) processes to that data. They do endless backups, disaster recovery (DR) replications, snapshots, copies and cloning to make sure they can do business the way they did when the data was dynamic--even though it isn't dynamic anymore. This means they run out of space, bandwidth, archive space, tape and so on. Then they have to go back to the boss and ask for more money.
But not you. You realize that when data passes between stages you should adjust your processes accordingly. By doing so, you don't need to get more expensive gear, bandwidth or people. You simply migrate the data based on your own subjective, intelligent rules. You alter your backup/DR/test planning so it creates fewer copies of the same data and breathe much easier in general. You add helpful technologies, such as data dedupe, to further reduce the need to go back to the well. You're smart and popular. You, my friend, have mastered the art of common sense. Certificates are forthcoming.
This was first published in September 2008