This article can also be found in the Premium Editorial Download "Storage magazine: CDP 2.0: Finding success with the latest continuous data protection tools."
Download it now to read this article plus other related content.
A great example of this is the availability of administrative control. The service and its capabilities need to be sized for the environment. There are many consumer and small office/ home office (SOHO)-grade solutions that are attractively priced but not designed or packaged to fit the needs of larger companies. IT groups need to document minimum requirements. Some vendors won't offer capabilities such as departmental capacity limits, file/folder exclusions by user, advanced policy settings, centralized encryption key management or centralized admin. Here's a short list of factors to consider besides price when looking at SaaS options:
Retention settings and restores: Best practices suggest saving a copy of data for a specified period of time or for a specific number of versions. This way, data can be restored from any backup point within the period specified or from a previous version.
Backup SaaS solutions may provide one or both of these retention settings. Less-robust solutions don't offer a lot of flexibility, as they often retain data for only 30 days. More mature backup SaaS solutions may offer 30-day, one-year and seven-year retention period settings that allow organizations to conform to strict compliance guidelines. The most sophisticated services allow custom retention plans and/or archiving capabilities.
When it comes to online restores, some companies
| push files back to the original server online by default. Others allow you to choose the original location or an alternate one. On-demand restores allow administrators to have the ability to initiate restores at any time. These are all options to investigate and consider.
Recovery assurance: Service-level agreements (SLAs) should be of particular concern when an organization is handing over its data protection reins to a third party. SLAs, if available, will vary from one service to the next. Most vendors have fail-safe mechanisms in place to ensure access to the service to recover data--excluding factors that are beyond the vendor's control, such as sufficient bandwidth. More sophisticated vendors replicate data to geographically dispersed data centers to thwart service disruption in the event of a regional outage or disaster.
Because most recovery events involve only a few files at a time, the expected recovery-time service level is based on bandwidth and file size(s) for online restores. However, you should ask about expectations for full recoveries. Is there an expectation that the entire data set will be streamed over a WAN connection? And how quickly can the last full backup be assembled when using capacity optimization techniques such as data deduplication or redundancy techniques such as distributed RAID?
Sometimes you'll hear "We'll restore your data as fast as your Internet connection will allow," while more experienced providers will have computed the "speed of connection vs. volume of data to transfer" and may have realized that it's faster to drive the data cross-country than perform an online full restore. The latter typically offer some type of rapid-recovery service involving a portable disk device and third-party transportation. Turnaround time is dependent upon the amount of data recovered, the time of day the request is received and the method of shipment. Price can be affected by the cost of travel and staff time.
This was first published in October 2008