The Problem with Gold-Only Data Protection Service Levels

A recent Storage Switzerland webinar dealt with the impact of the massive increase in the number of applications that data centers support on the data protection architecture. How does IT apply the right level of protection, set service levels, for each of these applications? Setting service levels was hard enough when IT only had to worry about a few mission-critical applications and then a few more business-critical applications. Today, it is not uncommon for IT to be responsible for dozens of applications.

One of the webinar attendees suggested that the easiest way to deal with the problem is to provide the same level of protection to all applications, essentially creating a “gold” service level for all applications. The advantage of a “gold for everyone” approach is that IT doesn’t need to try to classify each application. The downside is the gold-only approach is expensive.

The Challenges of Creating a Gold-Only Service Level Backup

A gold service level typically means near-continuous backups and rapid recoveries. Both objectives are achievable, at a price. First, near-continuous backups mean using block-level backups, backup software-controlled snapshots or replication. Part of the challenge with frequent backups is capturing a clean copy of data, which means quiescing the application. The problem is that quiescing forces the application to flush its memory caches and not use memory for cache until the backup is complete. Although these various backup techniques work quickly, they do impact application performance.

Each of these “fast” techniques also has downsides. Block level incremental backups work by creating a golden master and then linking that master to incremental backups as data changes. During recovery the backup software has to “stitch” the golden master and multiple incremental jobs together to make sure it recovers the right copy of data. Most software limits the number of incremental backups it supports, in addition to the golden master. Many solutions require a consolidation job to update the golden master after storing a certain number of incremental backups.

Backup software that creates its copies by controlling a storage system’s snapshots doesn’t always get around the consolidation job problem of block-level incremental backup because most backup solutions still do a block level incremental of the snapshot volume. Some backup software can do a traditional incremental backup of the snapshot volume, but at some point, most IT planners want a complete consolidated full instead of hundreds of incremental jobs. Another problem with backup software-controlled snapshots is most backup software solutions don’t support every storage system on the market, most only support a few. Considering that most data centers have five or more storage system vendors in use, the chances of the backup software supporting all of them is rare.

Replication, a bit for bit copy of data from a primary storage system to another storage system, seems like a perfect solution. The problem is that most vendors require that replication be from and to like systems, which makes them very expensive. Also, these replication products still have to get a clean copy of data and require the same database quiescing as other solutions.

The Recovery Challenges of a Gold-Only Service Level

Rapid recovery for all applications is also possible, but it is again expensive. Most block-level incremental solutions also have a boot from backup capability. With this technique, the backup software instantiates the recovery volume on the backup device, shaving network transfer time off the recovery process. The problem is the device which the software instantiates the recovery volume on needs to provide performance similar to the production system. If primary storage is all-flash and the recovery volume is on a compressed, deduplicated, hard drive based storage system, then the performance impact is so significant that the users deem the recovery volume unusable.

Recovery from a snapshot of the application is only suitable for situations where the application has corrupted data, or the primary server has failed. If the storage system or the data center itself becomes unavailable, then the snapshot that is on that storage and in that data center is also unavailable, unless it is replicated out to a secondary site. Replication is again very viable but also very expensive from a rapid recovery standpoint.

IT Needs Multi-Grade Service Levels

All of the above techniques are required to deliver gold service levels, but they do add significant expense and some complexity. If the organization tries to apply these techniques to all the applications in the data center the cost of data protection may consume the entire IT budget. The cost to create silver and bronze services levels is significantly lower than the cost of gold service levels and for the right applications meets the data protection and recovery demands they require. IT needs to be judicious in its application of gold-service levels, especially in the face of the rapidly growing number of applications they are required to support.

In our next blog, we’ll discuss what silver and bronze service levels look like in the modern data protection architecture, and our last blog calculates the ROI of a multi-level service level strategy.

Sign up for our Newsletter. Get updates on our latest articles and webinars, plus EXCLUSIVE subscriber only content.

George Crump is the Chief Marketing Officer at VergeIO, the leader in Ultraconverged Infrastructure. Prior to VergeIO he was Chief Product Strategist at StorONE. Before assuming roles with innovative technology vendors, George spent almost 14 years as the founder and lead analyst at Storage Switzerland. In his spare time, he continues to write blogs on Storage Switzerland to educate IT professionals on all aspects of data center storage. He is the primary contributor to Storage Switzerland and is a heavily sought-after public speaker. With over 30 years of experience designing storage solutions for data centers across the US, he has seen the birth of such technologies as RAID, NAS, SAN, Virtualization, Cloud, and Enterprise Flash. Before founding Storage Switzerland, he was CTO at one of the nation's largest storage integrators, where he was in charge of technology testing, integration, and product selection.

Tagged with: , , , , , , , , , ,
Posted in Blog

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 25,553 other subscribers
Blog Stats
%d bloggers like this: