Library maintenance: reducing your TCO.

It used to be that IT departments operated reactively, adding hardware and infrastructure in response to often unanticipated system overload or increased data growth. Innovations in software have led to better ways to use data that were unimaginable before networked computing. This has naturally led to a proliferation of computers, applications, infrastructure demands, data, data gathering and data parsing tools. Much of this explosion of technology could not have been managed in any other way than reactively, given the rapid changes now assumed to be the norm in information-rich business environments.

This newly established norm, however, has given businesses a chance to evaluate costs that were once as unanticipated as the technology explosion that drove data growth. As IT departments continue to accommodate data's growth and value, businesses are learning to assess IT as a business unit, planning more carefully for IT expenses.

With IT increasingly adopting a business-driven model, the storage industry is evolving by supplying a business case for products that solve real business problems. Automated tape library manufacturers, for example, are putting as much thought and effort into the business case for their machines as they put into the development of new technology. They now address issues such as cost savings, risk reduction and investment protection in addressing the total cost of ownership (TCO).

When analyzing the TCO of automated tape libraries, costs are mostly related to hardware acquisition, operation, and associated staffing requirements. These aspects of TCO are important, but one significant cost is often overlooked: maintenance. Library maintenance seems to not be specifically addressed by many chief financial officers in their analysis of TCO because it falls into all three of the financial yard-stick categories mentioned above:

Cost of acquisition: Along with initial hardware costs, the cost of an appropriate technical support contract should be considered

Cost of operation: Along with the costs of installation, system integration, floor space and power requirements, there are additional costs required to reduce library failures and service calls, as well as expensive downtime 

Cost of staffing: Along with the staffing required to handle daily operations, the cost of staff devoted to maintenance needs to be considered

Over the life of the library, maintenance can sometimes exceed the purchase price. However, more and more library manufacturers are making an effort to change this. This article covers how the cost of library maintenance figures into TCO, as well as ways in which companies can reduce their TCO by carefully selecting the right library and available options.

Acquisition

The cost of acquisition includes more than just the cost of the hardware; it also includes technical support that meets the needs of the organization. Fortunately, technical support organizations have moved beyond the reactive support once only offered over the telephone and supplemented by field engineers. Library manufacturers and resellers are now offering a variety of support plans and more of them at increasingly affordable rates when their libraries are designed for partial self-maintenance and automated support.

On-site support is critical to many organizations, so it is important to have comprehensive on-site maintenance programs delivered by a highly qualified team of professionals. A qualified support team has years of industry experience in diagnosing and correcting problems with libraries, drives, operating systems, networks and software.

In addition to on-site service, other plans should allow data centers to host spare components and use data center staff for repairs. Many data centers value this option, since providing maintenance services themselves saves both money and time. Of course, self-maintenance requires a library design that lends itself to easy repair. For example, libraries that are composed of distinct, individualized modules facilitate diagnostics and troubleshooting and allow users to make their own repairs.

Automated support features--including, for example, a library's ability to proactively and intelligently monitor library functions and e-mail status and diagnostic information directly to support staff-are becoming increasingly valuable. Coming directly from the library, such features provide support staff with detailed information about any problems, as well as access to history and configuration data. This immediate access to important background information proactively ensures maximum uptime for the library.

Operation

Effective library performance is crucial to the success of IT operations, as well as to the company's overall business. What would a company do if any of its valuable data was lost or damaged? With increasing pressure to save data for longer periods and keep it from being edited, companies need to be absolutely positive that their data is stored accurately and securely. Maintenance ensures that the library always performs at its best. Usually this means preventative maintenance--making sure that all components operate without failure. Regular drive cleanings can be performed by data center staff, and library diagnostics can be performed as necessary. As a result, adjustments are made before problems occur because, after they occur, the costs multiply. The time and resources put into recovering lost or damaged data just aren't justified, in comparison.

Determining the TCO for protecting data without failures that affect day-to-day operations requires a look at the cost of downtime. In "How to Partner for Disaster Recovery" (Computer Technology Review, April 2003), Jason Buffington, director of Business Continuity for NSI Software, explains such losses are calculated by measuring the productivity impact if critical data becomes inaccessible. Simply take the cumulative hourly wage of all employees who cannot work without the data, and multiply it by the number of hours that the data is inaccessible. You must add to that number if the affected employees directly generate profit, such as sales. In that case, divide the company's average monthly gross revenue by the number of down hours for lost profitability.

Adding the frequency of outages also helps a company to understand loss of data in relation to TCO and stresses the importance of maintenance in reducing that cost. As mentioned previously, companies should seek out libraries that are designed with data center staff in mind for handling all maintenance tasks. Such an option provides immediate support, right when needed. Again, having this option available to IT staff requires a library with key components--such as power supplies, drives, interface modules and robotics--that are modularly designed. Imagine the cost savings with a self-maintenance program that allows replacement components to be stored on-site, where staff can access and replace them right away. Compare that cost to that of having to wait on a field engineer to arrive and replace a complicated part that can't be easily extricated from a non-modular library.

Labor

IT staff have their work cut out for them. Their regular duties include systems backup, systems monitoring, software updates and patches, hardware updates, performance optimization and application maintenance--and that's on a day when nothing goes wrong. When a failure occurs, staff needs to turn their attention to problem diagnosis and resolution, both of which can consume large amounts of time and energy.

In "It's All in the Delivery: Moore's and Murphy's Laws Have Redefined IT Support" (Storage Management Solutions, July/August 2002), Tom Berghoff, vice president of Business Operations for Worldwide Support Services at Sun Microsystems, writes that the never-ending development of new responsibilities is causing IT managers to turn to new resources and tools to solve their own problems.

Think about it: Automation was originally developed to save time and labor. So why should it bring additional requirements to maintain it? As to the level of labor required to maintain libraries, the most important considerations include the tools provided with them for regular maintenance. Even the most stable and reliable products require proactive maintenance and reactive attention. Therefore, the maintenance tools included with a library must provide unparalleled ease of use for IT staff.

One such tool is the library's user interface. By design, it should be intuitive and easy to navigate, as well as provide built-in diagnostics and troubleshooting tools. As in parts replacement, it is far faster for IT staff to be able to handle a problem on the spot; they don't want to wait for a field engineer to arrive and figure out the problem. The self-diagnosing tools provided in the user interface are even more useful if the interface is accessible remotely; that way, IT staff can get information and run tests from anywhere. Imagine the time and effort saved when system administrators can receive an automated alert from a library at any time of day, connect remotely to the system from wherever they are, and figure out the problem themselves--all in less time than it takes to call support and wait for a technician's help.

Reducing TCO Through Maintenance

By carefully examining the options available in libraries, in regards to maintenance, you can control the library's TCO. Simply keep in mind the level of technical support required by your organization, the reliability and modular design of the library itself and the proactive maintenance tools provided with the library.

www.spectralogic.com

Hossein ZiaShakeri is vice president of advanced engineering and Catherine DeGraaf is senior technical writer at Spectra Logic Corporation (Boulder, CO)