As Kermit’s whimsical ballad explained, green isn’t as easy as we might have thought. The complexities and traumas of virtualization and consolidation of the production environment do indeed provide “green” savings. However, there is another significant opportunity for green savings with significantly less production impact.
Data protection media is the oft neglected child of operations. Conventional backup operations produce anywhere from 10 to 200 or more tapes every evening. Typical default policies result in some recycling. However these typical policies also tend to include long-term retention of monthly, quarterly, or yearly backup copies, often for periods of seven years to forever. The net result can be an offsite archive of tens of thousands of forgotten tapes that languishes unseen, unrecognized, and often forgotten year after year after year. Each tape consumes its own carbon footprint that includes not only manufacturing but also subsequent green impact in transportation, power and cooling, and floor space needed for secured storage for a decade or more.
Greening Data Protection
So how can this backwater of IT be brought to management’s attention in order to receive its just share of funding and due recognition for the critical business insurance it provides? Today, just about every industry magazine read by the CxOs has an article or two, or three, on the importance of being green. This new environmental awareness is projected to provide market advantage and perhaps even political advantage by demonstrating not only corporate awareness but corporate active participation in reducing the carbon footprint. Closer to home, actively addressing consolidation and efficiency opportunities can lead not only to real savings in power, cooling, and floor space, but also savings in administration through a simpler environment. Being green can be a great lever for encouraging a new approach to data protection.
The emphasis on being green has already promoted several initiatives in the data center, but none as yet in protection media. Today we already see a surge of servers being virtualized, resulting in significant reduction in the numbers of physical servers requiring facilities and administration. This provides very significant savings as numbers of underutilized servers are moved to a virtual environment on a single physical machine. Storage, too, is being virtualized heterogeneously. State of the art now allows storage to be allocated across storage frame boundaries simplifying the administration while leveraging the ability to better utilize existing storage. Thin provisioning is providing significant additional advantages by allowing traditional over allocation (to reduce risk of running out of room) while still constraining the actual amount of physical storage needed by just-in-time additions as actual physical use approached thresholds.
These key initiatives provide highly visible green impact that result in real savings in hardware and associated power and floor space costs. Additional savings can also occur as administration needs are reduced to match the newly consolidated environment. These savings are achieved within the highly sensitive production environment and require great care and planning to ensure the leap forward does not result in a leap backward. Nonetheless, this is classical change management and should be well within the grasp of the mature organization with well tested change and release management.
An additional opportunity for green savings exists in the humble data protection environment. A typical organization can have tens of thousands of tapes held off site. The manufacture of each tape leaves a carbon footprint, and the retention of each tape requires classical conditioned and secured premises that consume power and cooling. Additionally, physically transporting tapes to and from the storage facility consumes gasoline and pollution. Taking action to reduce the number of tapes, their transport, and their retention can result not only in green savings, but also can provide additional advantages to the perceptions of the data center as a service provider. The process necessary to rationalize the use of tape media can result in closer alignment of IT to real business needs in data protection and the use of state-of-the-art technology to improve services. The focus on being green and reducing the environmental footprint can also provide significant benefits in cost efficiencies for the beleaguered data center. Being green can now also mean being appreciated.
What is the Current Situation?
In many organizations today, the typical data center diligently creates and manages several generations of incremental backup, often 30 days or more, and even more diligently, maintains full backups over extended periods, often indefinitely. These retention periods are either an unthinking default, “we’ve always done it that way,” or an attempt by the IT Citadel to ensure they are covered against reprisals “just in case” something gets lost. Although substantially invisible to the CxOs, this hidden vault of tape media covering perhaps the last 15 years is achieved at significant green impact.
This we typically find an offsite store of tens of thousands of tapes, many of which cannot be read because the media itself has aged beyond manufacturer specifications and consequently degraded, and even more also because the tape platform no longer exists (reel to reel, DLTx). Even if the platform does exist, does the software that created the backups still exist? Does the catalog exist to retrieve the data? Can the data be rendered by current applications? Is there in fact any justification for keeping all this data which is generally believed, falsely, to be actual information?
Then there is the issue of physical integrity. I am sure many of you will have seen the occasional operator carrying a stack of cartridges in hand, and after dropping one, they simply nudge it along with their foot until they reach the cardboard box in which tapes will be placed for pickup by offsite contractors. The danger does not end here. The box of tapes gets carried down to reception where the contractor picks them up, loads them into the truck, and off they go to a distant storage point. Each point along the way can damage the tape so that a recovery attempt in the future may be totally frustrated. Then the fingers are pointing, and the users march with burning torches on the crumbling IT Citadel.
Why We Need Sane Data Protection Policies
Given that recovery point objective (RPO) is often a de facto 24 hours, the naive newcomer may ask why it’s necessary to keep tapes for 15 years. This results because there is often a de facto and undocumented “worst case” RPO that supports recovery of lost files that are discovered as missing long after the actual event. This is the classic “just in case” protection strategy of the prudent IT citadel, particularly in the absence of any corporate policy. This is exacerbated by additional assumptions that are even less likely.
The assumption that backup copies can be used as de facto archives is common. While senior management dithers on determining appropriate record retention policies and associated compliance and funding issues, the organization feels safe because “our backups are retained for seven years.” This de facto archiving by use of backup copies occurs under three additional and equally unsubstantiated assumptions.
First, the assumption is that tape technology still exists to read the tape. Second, the assumption that the tape itself is readable after all these years is often proven false under the most embarrassing circumstances during attempted recovery. Third, the assumption that platform, operating system, and application software can be made available at the version needed to render the data recovered back it into information is even more unlikely. The fourth assumption is that the rendering of recovered data by an application will not breach compliance requirements of immutability.
Developing Sane Protection Policies
Developing sane data protection policies aligned to business drivers brings a new reality to what has been traditionally a default process, typically invented by IT because no one else was interested. Recovery time objectives, and recovery point objectives that define how quickly an application needs to be recovered and how much data loss can be tolerated, need to be set by business units under an overarching corporate policy placing this responsibility solely with the business units. This approach, when supported by tiers of service in the classical service provider model, allows business units to make a cost-conscious decision on the tier of protection service needed. Similar policies need to be set on data retention including not only the retention period but the desired retrieval time objective, should the data need to be retrieved from the archive.
These policies result in development of tiers of service to meet business needs in data protection and retention. With supportive reference architecture, state-of-the-art solutions, and documented standard processes, this new strategic approach can provide not only significant green savings in media, floor space, and transport costs, but also improves IT alignment with business units. This alignment drives down costs and increases end user satisfaction. Indeed, being green in data protection can have significant payback both beyond the corporation, within the corporation, and within the IT citadel itself.
The adoption of sane data protection policies (i.e. business driven) under the service provider model can lead to the elimination of tens of thousands of useless tapes and ensure the business unit’s needs for data protection, retention, and archiving are met with clearly defined and costly services. Indeed, as the policies and supportive processes are developed, and reference architectures designed, it becomes apparent that electronic vaulting to an external location can provide additional green benefits while at the same time improving data protection service levels. No more diesel trucks arriving and departing every day, no huge off site rented vaults holding unrecoverable data, but best of all, data protection is now real.
How Technology Can Help?
State-of-the-art addresses the basic rational for tape media and mitigates the need for this ancient media by encouraging a move to VTL (virtual tape libraries) where backup software unwittingly writes a tape image that is actually written unbeknownst to disk.
State of the art also encourages us to think beyond traditional data compression techniques and consider the benefits of de-duplication. Full backups will inevitably back up a high percentage of unchanged data compared to the last full backup. With an average change rate of around 15 percent, perhaps up to 80 percent of data from the last full backup will be unchanged. De-duplication can identify this and save large amounts of storage.
State of the art also encourages electronic vaulting where data is copied across the network to its data protection storage rather than being copied to a local data protection store (tape library) and then ejected for physical transport offsite by trucks.
Some advanced operations have now achieved the ability to electronically vault to their data protection site miles away and have purchased libraries of sufficient size that tape handling by humans is a rare event. Indeed, any human handing of tape media creates the risk of damage that can result in the data being unreadable. The less handling a tape cartridge receives the greater the chance that physical integrity is maintained.
While technology plays a critical role in impacting use of “green” resources and reducing costs, the real key is to develop policies at the organizational level that define the business rational for the protection and retention of data.
Only effective policies can dramatically reduce the number of tapes in use, the amount of data copied for protection, and the number of tapes held off site, while providing green savings and enhancing the organizations ability to recover data to business driven goals. Technology can help, but technology alone is not the answer to being green.
Dick Benton is the principal consultant for GlassHouse Technologies. Benton brings many years experience in the role of technology governance and disaster recovery consulting. Benton has written extensively on these subjects and has helped numerous clients to develop business aligned disaster recovery strategies. Benton can be contacted at email@example.com.
"Appeared in DRJ's Winter 2009 Issue"