My previous post Is a Copy a Backup? was written after I got to thinking about a backup services catalog. What is a backup services catalog? Basically it is two things.
First, it is an admission that we have a lot more choices in terms of what technologies to use to provision backup services than ever before. 10 years ago, the backup technologies was really backup technology: tape. It was pretty easy to design a such a catalog, because you really only had one choice: tape. And it didn't really matter how rich, thin, or beautiful you were, you still only had one choice: tape.
Second, it is a recognition than different applications have different levels of importance to the business, and that given we have an array of technology choices with which to backup them up and recover them, we should probably choose different technologies for the different applications. Let me put that a different way: your SAP/Oracle application which runs AP, AR, and inventory is probably way more important than a miscellaneous SQL server some user threw up for ad hoc reporting, or an edge web server that is one of a few dozen others just like it.
So if you have different tiers of applications, shouldn't you have different tiers of backup and recovery? You bet you should.
The first table below attempts to describe what the various technology choices are today in the world of backup and recovery. They are listed on the top row. The columns detail what I think are the most important characteristics of these technologies: what Recovery Point Objective (RPO) can they provide? What Recovery Time Objective (RTO) can they meet? How long is data normally retained on them? What level of performance and reliability do they provide?
If we glance at the table, we can see that, very roughly speaking, the technologies are ranked in terms of capability left to right: most technical capabilities decline as we move to the right. Incidentally, costs per TB tend to decline in the same order (meaning that CDP technologies carry a higher cost per TB than Archive technologies for example).
I have not however provided firm costs per TB of each of these. I didn't give specific pricing because clearly that would depend on what vendor implementation you decide to go with, and how effective your negotiations are with them. I also didn't give specific prices because this is actually a lot harder question to answer than it might first appear. Cost of acquisition may be easy to figure out, but cost is really only interesting in this contect if it is Total Cost of Ownership (TCO).
So I don't really care that virtual tape costs me $5/TB to buy (and for arguments sake, lets stipulate that includes 3 years of maintenance). What I really care about is that $5 plus my costs to managed, cool, provision bandwidth, save me in other areas, reduce my risk, and so on. Everything that goes into a comprehensive TCO study.
So I am going to stipulate the technologies are roughly positioned in terms of cost (although I do think the last 3 are very close in cost per TB, and that tape really only appears to be cheaper when you are talking about using it for monthly archival images that you intend to retain for years... any more demanding requirement in terms of performance or capacity, and it is likely as expensive or more so than deduplicated disk.)
I am also going to stipulate that the average business probably can identify 3 tiers of applications. I know that some businesses have more (five, or six) and some have less. However, to avoid a bunch of complexity, lets just assume we have 3. Tier one applications would likely include SAP/Oracle production applications and email. Tier 2 might be large file servers. Tier 3 might be edge systems, small database servers, and so on. My contention is that it is probably pretty obvious to most people what applications should go in what tiers. It may be really hard to justify those intuitions and suspicions, especially if the application owners are cranky and insist that they "need" a higher tier of service. But most of the time if you just have 3 tiers, what goes where is pretty obvious.
Given all that, the next chart attempts to describe what I think are probably a cost justifiable set of backup and recovery services for the average enterprise and the three tiers of applications. Again, your mileage may vary as you assign specific TCOs to each of the technical offerings. But I think that many businesses would likely land on a model pretty similar to this.
What does the chart mean? The top row is the three tiers of application service. The left column contains the different technical choices. And the intersections are how long data would be retained within each technology before being disposed of (or, in some cases migrated to the next offering). That is to say that a Tier 1 application would get 7 days of CDP, have backups retained on VTL for 7 days, on deduplicated disk for 6 months, and long term archives retained on either disk or tape for 7 years.
This would constitute a service catalog for backup and recovery. Assuming that you did the TCO exercise and assigned a specific $/TB to each technical offering, you could then offer the business 3 tiers of backup and recovery service to match the three tiers of application service. And you would be able to say to a given business unit or application owner: "Here are your costs for each of the tiers of service. Pick the one that meets your business requirements, and that you are prepared to pay for."
Another useful outcome of this exercise, incidentally, is that you can force business and technical users to think about how much services cost. In my experience, businesses universally want the very best service you can provide them, unless they are forced to pay for it. Everybody wants gold service with lots of snaps/clones, lots of backup copies, kept forever, on fast disk, and replicated everywhere. Until they figure out the price tag. Then... OK, maybe just regular backup to deduplicated disk and remotely replicated is good enough.
(An amusing anecdote: right after 9/11 an infrastructure company asked us to determine how much it would be to provide a complete disaster recovery capability for their data center. After a couple of months of study, we concluded that it would be $12m to $15m in infrastructure only. Not including software, services, skills, or time. The board of directors took one look at the figure and told them to make do with what they had. The same board that asked for full DR months earlier. Everybody wants a Ferrari. Not everybody wants to pay for a Ferrari.)
All this is a pretty grotesque simplification of the actual process. In particular, determining the TCO of the different technical offerings is a non-trivial task. And getting business owners to admit that their applications may not actually be tier 1 is often no easy task either. But by offering them a rational set of choices, with reasonable costs associated with each, you can deliver the quality of service that makes sense for any given application.
And hopefully you can deliver a better service overall: at roughly the same cost point as it would be to deliver a single, generic backup and recovery service to every application, it should be possible to offer a higher quality service to those applications which require it, and a lower quality of service to those that do not.