Unitrends recently introduced the Unitrends Recovery-943. This appliance is the next generation flagship of the Unitrends Recovery-series of backup appliances. We began this series of blog posts by discussing why backup appliances are experiencing such rapid demand and adoption, and have continued the discussion over the next few posts by addressing the challenges facing enterprise-class data protection backup appliance vendors. We will then continue the discussion by exploring the Recovery-943 architecture and the specific responses that the designers of the Recovery-943 had to each of the challenges discussed in this post.
I recently bought my family some very upscale Tumi™ luggage. One piece is a beautiful hard case carry-on. Unfortunately, I can’t use it – I have a tendency to pack more than the hard case carry-on can hold and I end up using a soft case carry on instead that holds more (and was a lot less expensive.) What I value is the amount I can carry. Others in my family value the superior aesthetics of the hard case carry-on more than the amount of stuff that fits in it.
As more companies have sprung up to take advantage of the new technology of solid state drives, an interesting argument has been proposed by the vendors of these products. An argument is made by most of these vendors that affordability should be measured not by traditional measures such as dollars per terabyte but by new measures such as dollars per I/O operation over some unit time. It’s an interesting position – compelling in some compute critical environments while utterly erroneous in other environments.
Affordability in data protection environments is an example of an environment in which affordability is still measured on your “backup bang for the buck” – the overall cost of capital and operational expense associated with backup functionality and retention available for each dollar spent.
Great designers of backup appliances are constantly aware that they must balance affordability with superior functional capability and performance. In other words, the most capable and highest performance backup appliance in the world doesn’t do much good if it’s built entirely out of solid state devices and thus the price per terabyte is so high than no one can afford it.
When my daughter was in her mid-teens, one of the things she said that I absolutely despised was when she responded with “whatever” when I asked her a question. I still don’t like it when people use that phrase – it implies to me an abdication of not only responsibility but of opinion that I typically just want to verbally fight through to understand the person to whom I’m speaking.
Paradoxically, I love virtualization – even though I’ve always felt that with the abstraction it affords it was the computer industry’s way of saying “whatever” when confronted with technical challenges. And sure enough, with each passing year, we as an industry and those of us in IT learn that “virtualization” isn’t an answer – it’s a technique for coming to a better answer.
Backup appliances, and backup software, of course must use various techniques to protect virtualized environments. The arguments regarding HOS (Host Operating System)-level and GOS (Guest Operating System)-level virtualization data protection techniques are beyond the scope of this paper ; however, it’s worth noting that any serious data protection product should be able to flexibly adapt and handle multiple virtualization use cases.
An entirely different challenge for backup appliances is incorporating virtualization to accelerate recovery, in other words, to minimize what is typically called in our industry the RTO (Recovery Time Objective). The use of this technology, which is called virtualization failover, can significantly impact the operations of a backup appliance.
Networking and Connectivity
Henry Ford is famous for saying “Any customer can have a car painted any color that he wants as long as it is black.” As can be seen by simply driving down the road today, that particular philosophy didn’t last long as it came crashing headlong into the reality that customers wanted choice.
Today there are several backup appliance vendors new to the space who offer any type of backup appliance that you want – as long as it’s the single model that they offer.
But beyond this foolishness, there’s a tendency across the backup appliance space to limit the capabilities associated with networking and connectivity. The overt rationalization is that appliances should be simple – and that’s absolutely true. But the underlying reality is that vendors do this in order to limit their internal configurations. This makes short-term sense in terms of these vendors saving money; however, it doesn’t respond to the incredible importance of adapting and optimizing networking and connectivity with respect to the incredible diversity of today’s agile IT infrastructure.
Replication and Archiving
When I was a kid, I thought one of the greatest things about Star Trek was the transporter. The ability to physically move a person (or a thing) from one place to another instantaneously was really “fascinating.” I particularly liked the episode where there was a malfunction and the “copy” of the person being transported wasn’t destroyed – that was the first time I thought about the fact that transporters worked by copying rather than by moving the individual.
Replication and archiving are important features for modern data protection. Replication does exactly what it sounds like – it replicates, or copies, data from one place to another. It does this electronically by copying data to what is typically a distant location. Replication offers disaster recovery at the lowest operational cost – but takes WAN bandwidth and also takes capital expense in the form of equipment at the other end of the replication. Advanced replication takes source level deduplication, compression, and encryption as key technologies to make it possible to protect large amounts of data over limited WAN lines effectively and safely.
Archiving creates a copy of a backup on either rotational or fixed media. Rotational media are devices such as disk drives (connected via USB or eSATA, for example) or tape drives (connected via LVD SCSI or SAS (Serial Attached SCSI.) Rotational archiving is sometimes called “sneaker net” because you’re “walking” your copied backups (or archives) to safety. Safety here can mean the trunk of your car each night, shipped via truck to a mine underneath a mountain, and the like.
Fixed archiving is typically performed for nearline retention extension from the backup appliance. A typical target is a SAN or a NAS.
One thing that’s important to realize about archiving is that the boundaries between rotational and fixed tend to blur. You can archive to a small NAS and then take it off site. You can use disk and tape to extend retention. And there’s even blurring between archiving and replication with the new cloud-based NAS and SAN products from companies such as TwinStrata, AWS’s integrated cloud gateway, and even Dropbox.
Before we begin our blog post, which will discuss appliance architecture in detail, are there any challenges facing enterprise-class data protection backup appliance vendors that I’ve failed to mention? If so, leave me a comment.