Managing a Tape Environment Is No Longer an Oxymoron

Managing the software and hardware components of an IT infrastructure, such as in a data center, has become an increasingly complex and difficult challenge. Yet meeting service levels and cost requirements revolves around the ability to manage efficiently (do things right) and effectively (do the right things).

Almost all vendors tout how well their software tools can help ease the burden of management. While that is desirable, simplified management is not the only answer to the complexity challenge. A more important question to consider is whether administrators have the necessary information and analytical capabilities to manage effectively -- if at all. For example, IBM just acquired Exeros to enable it to help customers automate and accelerate understanding of data and data relationships. The rationale that IBM expressed for the deal is, "You can't manage what you can't understand." That builds upon the age-old philosophy: "You can't manage what you can't measure."

Vendors are trying to fill in as many data center management gaps as possible, especially in an increasingly virtualized world where the focus is on creating a new generation of IT infrastructures. Storage vendors recognize this, as well. For example, NetApp bought Onaro to acquire SANscreen, an approach to extending data center automation to storage. EMC acquired WysDM for data-protection management capabilities that provides past, present and future insights into existing backup environments.

But what happens when the information that you need to manage is not actually available? For example, all the information that is needed to manage a storage area network (SAN) is not natively available. HBAs, switches and disks are all involved in SAN I/O, but it is not their job to collect that information. Adding in specific hardware devices to be able to collect SAN traffic data enables the ability to audit and report on data traffic and trends. A company called Virtual Instruments provides that capability, which can lead to quickly identifying performance problems and behavior anomalies, as well as being able to point out potential issues that are on the verge of exceeding "best practices" thresholds, and moreover, identify potential cost savings.

That is great for SANs, but what about the tape automation environment? Tape automation is not the most glamorous part of a data center (as a user never sees it directly), but it is a major part of its bread and butter. If a data center were a starship, Scotty would be very comfortable with the tape environment as part of his engineering responsibilities.

In fact, many engineering disciplines come together in a tape environment, including mechanical and electrical engineering. A tape automation environment has three basic elements:

Tape automation -- typically a tape library built around robotic arms whose movements are a triumph of mechanical engineering.

Tape drives -- which ingest tape cartridges, control the movement of tape along tightly defined paths and expel full tape cartridges to be moved by the library robot to an available storage slot.

Tape cartridges -- these compact and secure huge amounts of data in a small space using sophisticated technology.

The interplay of these three elements is interesting to watch, resembling a dance routine running at high speed (such as the movement of a robotic arm). However, natively these elements do not capture much information that can be used in their management. In fact, management of the tape environment is likely to be reactive instead of proactive, such as repairing a tape drive that refuses to ingest or eject a tape or replacing a tape cartridge that cannot be read (a very bad thing if it is part of a major recovery process).

To our way of thinking, tape environment management should be much more than reactive failure analysis. Proactive management can be constructed around answering a number of critical questions:

-- Can alerts that are the results of degraded performance or serious error conditions be reviewed properly so that corrective action can be taken? If not, backups may not complete correctly and critical restores could fail.

-- Can the interaction between a drive and tape cartridges be examined to determine the root cause of a problem? If so, this can resolve a major source of administrative headaches.

-- Are tape drives performing efficiently? A simple question whose answer is critical, since inefficient performance could lead to unnecessary high failure rates.

-- Are tape drives being utilized effectively? Overused tape drives can have high failure rates while underused drives result in a waste of scarce IT budget dollars.

-- Can an already written tape be read correctly? Another simple question whose answer is crucial for ensuring recovery processes.

Crossroads Systems has solved the management problem for tape automation environments through the introduction of its ReadVerify Appliance (RVA) and supporting services. While Quantum has some very capable features for tape environment management via its Quantum Vision management software, the company focuses on its own, homogenous tape environments. In contrast, Crossroads can support any Fibre Channel-based tape environment (which includes open systems, as well as IBM's mainframe-based tape environment). Note also that Crossroads' approach focuses on the management of the tape storage devices themselves, rather than data management or information management.

The Crossroads approach is based upon three general principles:

1 -- An RVA is installed in the tape environment outside of the data path, collecting data that otherwise would not be available for analysis.

2 -- Crossroads' monitoring and reporting solutions not only have the necessary analytical and visual presentation capabilities but can draw upon one or more of the following sources of information -- a historical (i.e. longitudinal) database of the customer's own information, a database of vendor recommendations (such as what should be the appropriate duty cycle for a tape drive), and an anonymous database of what other customers have experienced that can be used for benchmarking.

3 -- Giving the customer an option of getting information through one or more Crossroads' services, which frees them up for the actual decision-making and action-taking rather than having to worry about the process of monitoring and reporting the information.

The customer can choose to purchase the RVA directly or work with a Crossroads partner as part of a service. Crossroads' RVA is a 1U high (i.e., 1.75 inches) rack-mountable "box" that sits non-disruptively in the tape environment data path. Naturally, the RVA ties into the LAN environment in so it can be managed and transmit its findings, such as email alerts, as necessary.

A customer can choose to work with a Crossroads partner to use one or more of the following tape environment services:

Library Monitoring and Alerts -- this service provides daily reports on all triggered events, such as lost device communications or serious error conditions. However, the service can also report on less critical, but still important alerts -- described as cautionary, concern or simply informational conditions -- that describe poor drive efficiency or high device load counts, which may have a long-term impact on systems' costs, capital expenditures or successful data recovery.

Tape Environment Site Analysis -- a survey report is generated after, say, a 90-day analysis of a site's tape environment, offering a comprehensive analysis of its condition. This includes drive-tape cartridge error analysis of the interaction between drives and tape media, overall tape performance over the analysis period, and whether or not the utilization of the tape drives are reasonable or not. The report uses clear, easy-to-view, sophisticated visual approaches (such as spider charts), as well as reference information generally not be available to the customer, so the site analysis is best delivered as a service.

Tape Verification -- using an otherwise empty tape drive, the RVA appliance can verify that the information written on tapes is still correct. The resulting reports show metrics of all tapes verified, which is useful for providing clearly documented records for regulatory compliance (think Sarbanes-Oxley, where restoring financial data is necessary to avoid big problems) and for Service Level Agreements. Overall, IT can rest easier knowing that tapes are good in case a recovery is necessary.

Typically, I avoid injecting my personal experience in my observations. However, many years ago, I ran a Fortune 500 data center that had a mixture of mainframes and minicomputers (now called open systems servers). On many occasions, I needed more information about the non-mainframe environment, and I kept asking my managers why I couldn't have the same information that was available for the mainframes. Since they had no control over the operating system of the non-mainframe systems, my requests could not be met. But I never give up and so, even though I am no longer managing a data center, I want IT to have the tools and information that they need.

And that includes tape. My data center had to manage 10,000 tapes even then. And we had one device we called the tape drive of last resort in the mainframe environment that was used to read what other tape drives might not. This is the reason I find the Crossroads solution interesting enough to bring to your attention. Tape is not dead (and not likely to die as soon as some pundits would have it), but while it is alive, it deserves to be managed well. Crossroads combination of RVA along with its three tape environment services should provide enterprises the tools they need to better manage and maintain their tape environments.