The disconnect between how quickly and efficiently end users think their IT department can back up and recover data and the IT department’s actual ability to deliver on these expectations can be substantial. Too often, IT departments are not equipped to recover data nearly as fast as end users expect, and they may not even have the data available to recover. In this fourth installment of DCIG’s interview with STORServer President Bill Smoldt, he explains why misconceptions about backup persist and what backup paradigms must change for the benefit of everyone.
Jerome: Backup is a core IT service in every organization. However, it is not necessarily a task that IT gets rewarded for focusing on or doing well even though recovery is a task that end users expect IT to perform quickly and flawlessly. Why the disconnect?
Bill: The way I have always looked at backup is that it should be boring, but you certainly do not want exciting restores. You do want to be excited about them, but you do not want them to be exciting. Historically, restores have been exciting and RGEs (Resume Generating Events.)
STORServer really helps a customer comprehend their backup environment. A prevailing presumption exists with endusers believing that absolutely everything is being backed up and can be restored. That presumption is corrected only during the rare restore of older data. They believe that every file is being backed up all the time and that they can recover anything from any time.
One of the examples I like to use to dispel that view is a typical grandfather, father, son backup paradigm where a customer keeps a weekly backup for some number of months, a monthly for some number of months (or years), and then a yearly after that. In reality, a file that gets created and then deleted during any of those periods of time simply may not be there anymore.
In my example, Joe User asks for a file that was deleted two years ago. Bob Admin restores the file from a three year old full (after a bit of a struggle finding the tape). Joe User sends an email to Steve CIO stating how wonderful Bob Admin is and how Bob is underpaid.
The example concludes when Joe User returns and asks for a file that was deleted last month. Bob Admin works over the weekend but finally has to inform Joe on Monday that the file had been created and deleted during the work week, and it can’t be restored. Joe User sends a flaming email to the CIO suggesting that Bob doesn’t actually have a clue what he’s doing and should work in another company.
The only data that got backed up is the data that happened to be on the system at the end of a particular time period. To the contrary in our paradigm of doing backups, STORServer can guarantee that for a given period of time, every file that got created and deleted and backed up is capable of being restored and an entire directory can be restored faithfully back to a particular point in time. That shows up in the interface without the admin having to guess or spin through tapes. The graphical interface defines the SLA that the backup administrator is delivering to his data stakeholders, no guesswork. And that data can be restored without operator intervention or recalling tapes from the vault company.
It is important to know not just the technology, but exactly what is going on in the backups. That can get very confusing with all of the newer paradigms for doing backups and restores. It’s important to have those new capabilities such as take a snapshot, replicate the deduplicated snapshot, and then keep the snapshots for some period of time. But you have to know how that fits into the overall backup scheme and it has to be comprehensible by the backup. Otherwise you end up with backup data sprawl and wasted resources.
One task that STORServer does is educate backup administrators on more than just the paradigm they are using and what they can actually deliver to their end-users. It also helps educate their end-users on aligning their expectations with service level agreements (SLAs). STORServer strives to ensure that there is never a surprise as far as what historical data is available and what is not.
All of that backup comes with a cost. All of this has to do with matching what the end user expects with the budget of the IT department. STORServer has so many rich ways of delivering that it now can deliver just about any SLA that somebody wants within a reasonable budget. But, they do have to understand that, as with all things in IT, none of this is free.
Part I of this interview series covers why large organizations can get up and running faster using STORServer’s backup appliances as they have the knowledge and confidence that they can backup data on any file or operating system.
In Part II of this interview series, Bill Smoldt provides some insight into how backup appliances have evolved over the last decade as well as the features they must offer to stand the test of time.
In Part III of this interview series, Bill shares why the cloud, deduplication and replication are the new must-have features on backup appliances.
In Part V of this interview series, we discuss sizing backup appliances and the growing popularity of virtual appliances.