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 will continue the discussion in the next few posts by addressing the challenges facing enterprise-class data protection backup appliance vendors. We will then discuss 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.

Failure and Data Loss

Imagine you were designing an airplane that would be used for transoceanic flight. If you design the airplane with one engine, you minimize failure. There are fewer things to go wrong, right? And there are other advantages. You can sell the airplane you create for less money and still make more profit. And the airline service that buys it can thus charge less for tickets and make it cheaper for people to fly to visit their families and do business across the ocean.

Minimizing failure seems like a really great idea – until you think about the consequence of failure. If you design your airplane with two engines and insure that the airplane can successfully fly with only a single engine, you double the risk of single engine failure on any particular flight but you dramatically lower the risk of the airplane crashing into the ocean.

Backup appliance designers live in a world in which failure is a constant threat which must be avoided when possible but always planned for and handled regardless. Backup appliance designers carefully use an assortment of advanced techniques to handle failure. A backup appliance overall and each component in a backup appliance is evaluated for its potential for failure using tools such as MTBF (Mean Time Between Failure) and AFR (Annualized Failure Rate.)

Failure is acceptable in a backup appliance; however, data loss is not. A backup appliance designer uses techniques such as MTTDL (Mean Time To Data Loss) to provide an evaluation framework for measuring data loss risk. While these types of techniques are invaluable, the most important tool for a backup appliance designer is constant, agonizing, white-knuckled, ulcer-inducing, teeth-grinding paranoia. The mantra of a backup appliance designer is not “what can go wrong might go wrong”; instead it’s “what can go wrong will go wrong.”


As products have increasingly grown more complex, it’s been more and more difficult for users of those devices to figure out what’s broken when something goes wrong. Even well-trained support personnel at companies selling products often have a difficult time diagnosing issues. And as everyone can attest, finding well-trained support personnel at companies is often impossible – quite often when I call support for a product I own I feel that I know more about it than the person paid to help me resolve what’s broken.

One advantage that vertical integration brings in terms of products is that it makes it easier for designers of products to build into the product itself enhanced diagnosability. This only makes sense, right? One key trade-off of any all-in-one integrated product is enhanced diagnosability at the expense of a more flexible (and complex) potential ecosystem in which it operates.

Apple figured this out long ago and has a design philosophy around the tight control of its offerings. From music players to computers, Apple offers less flexibility than its competitors but offers greater degrees of both usability and diagnosability.

Ecological Sustainability and Density

The “green” movement toward a more ecologically sustainable society means to business a new examination of its roles and responsibilities with respect to environmental responsibility. Vendors have flocked to label themselves as “green” from a marketing perspective; however, there’s an underlying reality associated with the fact that by reducing and optimizing carbon footprints and other ecologically damaging characteristics tends to translate to lower capital and operational expenditures for those businesses as well.

Higher density data centers provide businesses with a reduced cost of ownership when compared to low density data centers. As a result, whether your equipment is housed in your own data center or is colocated in multi-tenant data center, the best capital and operational efficiencies are gained with the highest density equipment. Density in this case refers to the compute and/or storage capability per “U” of rack space within the data center.

We will conclude this section in our next post. Have we missed any challenges you would like us to address?