Data Protection – What’s the Right Backup Strategy?

Monty Blight, Peak 10 -- August 1, 2006

Whenever I speak with executives about data protection, the conversation goes something like this.

Me: “How much data can you afford to lose?”

Them: “None.” The response usually comes with a confused expression that I would even ask such a question.

Me: “And what’s your tolerance for downtime?”

Them: “Zero. We can never be down.” Confusion has turned to disgust while I race out of the office to design the requested “zero downtime solution.”

I’m now back with the executive and my solution is sweet. It includes redundant hardware platforms, each fully capable of supporting a full load, housed in multiple geographically disbursed data centers with synchronous data replication and global load balancing. The CEO is clearly pleased until arriving at the final page, the one that details the “Investment” required. That’s when I’m tasked with designing a “minimal” downtime solution.

Data Protection is a daily battle between business requirements, technology capabilities and budget. Executives want a zero downtime solution because they are being told that data loss means that their business fails. And it could. The IT professional charged with delivering the solution craves realistic expectations, acceptable backup windows, scheduled maintenance, current tools, and sufficient headcount. And then there’s the budget. How much do you pay to protect against something that may never happen?

The good news is that the right solution is available. What that solution is depends on the real answers to those important questions. How much data can you lose and how long can you afford to be without it?

Before we explore the available options, I think it’s appropriate to review a few terms that often find there way into a discussion on data protection.

Storage: Storage in all its forms (solid state, disk, tape, etc.) is the physical location of your data. How storage is used is what defines it as production vs. backup. Production is intended for real-time data use and access.

Backup: A secondary batch of data redundant to the primary platform. In the event of a system failure, this data will be used to restore the primary platform back to full production.

Recovery Time Objective: The speed in which the production platform returns to full capability after a system failure.

Recovery Point Objective: The maximum amount of data lost (measured in time) in the event of a system failure.

Getting Started

The first step in data protection, as with most technical solutions, is to understand what problem you are trying to solve. Start with a business impact analysis. If a short amount of downtime results in significant financial loss, the previously described solution might be appropriate. It makes fiscal sense to spend thousands of dollars to protect millions. The reverse is also true. It makes no sense to spend thousands of dollars to protect hundreds. What if the business impact is not only measured in dollars? How much is a brand worth? Credibility? Reliability? Whatever the conclusion, the entire team needs to agree on (or at least understand) the goal before the Data Protection Strategy is implemented. Is the Recovery Time Objective 30 seconds, 30 minutes, or 30 hours? Can we afford to lose a day’s data or does it have to be 60 minutes or 60 seconds max?

With that determined, review the Data Protection options and then implement the strategy that matches Recovery Time Objective and Recovery Point Objective with the Budget.

Available Options

On-site Tape and disk backup provide the ability to copy the contents of all or a designated amount of data from its usual storage device to a tape or disk-based device.

Tape-Based Backup Benefits Include:

• Traditional and widely accepted media
• Required by some compliance standards
• Easy long-term retention and off-site storage
• After investment in initial library, additional media expense is minimal

Disk-Based Backup Benefits Include:

• Backups can be faster depending on the type of network
  
• Restores are almost always faster (no physical tapes to find and mount)
  
• Network backups provide superior disaster recovery protection
  
• Flexibility with backup source and recovery target

Remote backup is an online, disk-based storage and recovery method in which data is regularly backed-up to a remote device through a network connection. It relies on the company’s regular Internet so depending on data size, can be slow and may violate some compliance requirements. Remote backup solutions should include hardware, software, media, implementation, maintenance, remote operations, support, and monitoring

Remote Backup Benefits Include:

• Backup from anywhere and restore to any IP-based device
  
• Anytime restore availability
  
• Hardware can be located in multiple facilities

Companies can also use their hardware/software platform to replicate data between devices. This solution requires duplicate hardware, replication software and (depending on the solution) load balancing equipment. It tends to be more expensive but minimizes/eliminates downtime.

Replication Benefits Include:

• Synchronous replication eliminates data loss
  
• Quickest recovery solution – re-route all traffic to live site instead of data recovery
  
• Can eliminate geographic risk
  
• Easier maintenance on platforms because can transfer the load

In-House vs. Service Provider

Successful data backup can be accomplished in-house or through a number of service providers including data centers and managed storage companies. Implementing your own backup solution is the right choice if you have access to the necessary capital for the requisite hardware/software purchase, have backup expertise and surplus engineering time in-house, and have data in sufficient quantity to realize economies of scale. If that is not the case, the service provider route may be best.

Summary

The right data protection strategy is unique to each company based on the business impact (or not) of data loss relative to the budget to protect that data. It can be provided in-house or through a service provider. But no matter what you decide to do, please make me one promise. Make sure that your primary platform is secure and make sure that critical data is backed up and stored off-site.

---------------------------------------------

Peak 10 is the leading independent data center operator and managed services provider, delivering scalable, economical and reliable solutions for hosting and managing complex information technology infrastructure.. Monty Blight can be reached at (704) 264-1016 or by email at monty.blight@peak10.com.

Back