Unified Recovery Management #4 – Technology Choices
In my last entry, I explored some of the considerations that you should include in an overall data protection and recovery strategy, including matching the value of the data being protected to service level expectations such as Recovery Point Objectives (RPO), Recovery Time Objectives (RTO) and overall costs.
Today I’ll cover some of the many technology choices that are available to help you meet your objectives. As in previous installments in this blog, this adds another layer of complexity to the program – which technology do you use to meet which need? And at the end of the day how many different tools can you really manage effectively to meet the complex challenges of protecting your data?
There are many technologies and solutions available to help you solve the data protection problem, and new ones are continuing to be introduced. But each has its limitations, to the point where no one technology can address all of your specific data protection requirements.
Tape-based backup is probably still the most widely-used backup method in corporate and government environments. The challenges with tape have been well documented – lots of manual processes that can lead to errors and recovery problems; poor RPO and RTO performance; and the physical movement of tape cartridges that can create data security risks. For these reasons, many organizations have moved to a blend of disk and tape for backup, enabling faster and more frequent backups, and faster restores from disk, while moving backup data to tape in the background for longer-term retention.
Mirroring and replication are good technologies for system-level recovery and fail-over. However they can leave you with a big hole in your recovery strategy – the loss or corruption of individual files - since any loss will be immediately replicated to the backup system, leaving you with 2 bad copies of your data.
Continuous Data Protection, or CDP, takes the benefits of replication and adds in point-in-time recovery options. The problem with CDP is cost – it requires far more storage capacity than other solutions, and can strain network bandwidth as well.
All three of the above technologies are also susceptible to being unable to recover any files that are open at the time of the data loss incident, such as a system crash – although there are utilities available to help mitigate this issue.
Snapshots fix the open file issue by creating application-consistent time-based recovery points. It is necessary to pause or “quiesce” the application for a very short period of time to accomplish a snapshot, but it’s far faster than a tape backup because it only takes incremental changes over a much shorter period of time. Snapshots can be run very frequently, often many times per hour, to meet aggressive Recovery Point Objectives (RPO). Hardware-based snapshot technologies are not always ‘application-aware’, so the consistency (ability to fully recover) of open files can be a problem.
Disaster Recovery and Business Continuity (DR/BC) services are key focus for many organizations, especially given the growing number of natural and man-made threats to normal operations. Some companies handle it themselves, others contract it out. Either way, you’ll need to balance overall costs against benefits, matching the needs of individual applications and locations to the service levels provided.
Data deduplication is a much hyped technology that, depending on where it is applied, can reduce the amount of data that needs to be backed up and sent over the network, or reduce the amount of backup capacity required, or both. Most of the gains claimed by deduplication are in environments that perform weekly full backups that cause an enormous amount of duplication. Check out my blog series on data reduction to learn more about this important topic.
Virtual Tape is a relatively new entrant to the market, combining the best efficiencies of disk and tape, and adding in data deduplication to help meet cost per capacity goals. As a backend repository, a virtual tape library (VTL) does not replace data capture technologies such as backup, replication or archive, but can be an effective complement to them.
I added Reporting to the diagram above, only because you’ll want to have visibility into the functionality and performance of your data protection and recovery environment n order to provide the assurance that your strategy is effective and meeting the needs of the business.
Depending on the decisions you’ve made regarding protection policies, one, several, or many of these technologies may be needed to completely solve your data protection problem. Here are some questions to ask:
- How many different tools are you using to manage data protection and recovery?
- How many different admin tools do you have to use?
- Are they all located in one place, or scattered around the organization?
- Can one person have true visibility and control of the entire data protection and recovery infrastructure?
- How do you handle changes in the environment?
"The postings on this site are my own and don't necessarily represent IBM's positions, strategies or opinions."