Replication has always been one of those technologies that people run scared from. The main reason is usually cost. When you think of replication, your mind drifts to solutions that replicate from one disk frame to another using block level technology, or to products that replicate files from one host to another. The vision that likely comes to mind is one of very expensive broadband connections for these products to work properly. The fact is that today, there are a bunch of products on the market that got smarter with your data and bandwidth. Technologies like compression and data deduplication now factor in, reducing the actual amount of data that goes over the wire. The trend to follow is replication is now merging into data protection in such a way that the two become seamless. Here are my Top 5 reasons you need to take a fresh look at replication to enhance your overall Business Continuance strategy.
Number 5 – Hands Off
If your data is critical to the ongoing operations of your business and if you lost your data, you could potentially go out of business. Backing up data is the first layer of protection you have to implement to safe guard around potential issues that could cause data loss. For most companies, there are so many aspects of backing up and securing your data that involve human intervention. A couple of questions…
If you are using tape, who is responsible for making sure that the tapes are where they need to be?
Are the right sets of tapes in the library?
Do you even have a library or are you using a single drive and have to juggle tapes to get through one night of backup?
How about moving that set of tapes off site? Are the offsite tapes adequately protected?
Are they in the trunk of your car?
Do you send them to a data service that you spend a huge sum of money on?
Do you back up to disk locally and then say you’ll get to making a copy at some point?
Replication removes most if not all of these issues. Backing up your data to disk removes all of the tape issues above. If it is done efficiently (i.e. with compression and deduplication), you can store a lot of data in a small amount of space. Replication to another location, whether that location is across the campus, to a co-lo facility, to the cloud, or to your closet at home, automates the process of getting the data offsite in case of a disaster. It removes the human component from the equation and lets you run in a completely hands off approach to protecting your data. As long as you monitor it and have sufficient mechanisms in place to notify you of any issues, you can set it and walk away knowing your data is always there just in case…
Number 4 – Compliance / Legal Precedents / Corporate Governance
So not too long ago, we kept data around because we defined it as a policy internally. Something we call Corporate Governance. Nowadays, we have a large sum of regulations that we must live with pertaining to data retention. Today, at least, they don’t apply to everyone… Thankfully…
Retaining data on tape means keeping the same format or retaining the old drives and keeping them running as you go forward and move into larger capacity tapes and drives down the road. Therefore, managing tape for long periods of time cost money to vault it properly. Storing data on disk allows greater flexibility and more options for retrieval. Do you need to keep all of that data on disk for all that time, you ask? Well, what are we really talking about?
The legacy way of backing up offered you a weekly full and several incremental backups during the week. Each set was at least a full copy, more with the daily backups. If you had to take 1TB of data and back it up weekly, not counting the incremental backups, you would be baking up 52TB a year. You would not likely send each week offsite for 7 years, although some do, but let’s say monthly… that’s 12TB yearly, again not factoring in daily changes and data growth.
Newer products change the whole paradigm. They only perform one full backup and the rest is incremental, forever. Most offer the ability to export a point in time full so you can use the traditional methods of off-sighting data. Some have a better way.
Let’s say you purchase a lower tier of disk, cheap and deep, they call it. Now, let’s also say that the issues of replicating data such as bandwidth and cost were suddenly gone? And you could store 2, 3, 4, or more years’ worth of data in this offsite repository for a fraction of the total cost of a complete tape world… Now, combined with the automation, knowing you were able to access older data just as easily as the most recent, meeting the goals of retention, for whatever the reason, are met and the risk is mitigated.
Number 3 – Service Level Agreements
Sometimes you do things just because that is the way people want it. We call that an SLA. In IT, we live and die by the SLA. In most cases, when the critical server crashes and you fail to meet the SLA in a “reasonable” amount of time, you either learn to dance or dust off the ole’ resume. Replication gets your data to another location for safe keeping. It allows you to retrieve that data in the event of a failure at the primary location so catastrophic, your only recourse is going offsite.
An SLA might say “Service must resume from any failure within 24 hours”—pretty broad, but in most cases, very hard to achieve. You can’t get a server, in most cases, in that amount of time, let alone get it up and running with your data. Part of any recovery plan should be the process for getting hardware. This should be built into the SLAs. Once acquired, if you have a large quantity of data, and that data is sitting on several tapes that first need to come back from the vault, you could be looking at 24 hours or more just getting the restore running and even more time to complete.
With today’s data protection products that include some form of replication, this can be taken down to minutes. Meeting a tight SLA becomes more attainable at a fraction of the cost it once did in the past where you needed redundant everything to be able to get things back up quickly. This brings us to Number 2…
Number 2 – Cost
Let’s face it… money is tight. Everywhere you go, budgets are shrinking and not showing any signs of turning around anytime soon. The money is there for the important projects, of course, you just have to do your homework and due diligence.
Thing is, we spend more money when looking at the bigger picture of our current practices. Add this up for your company… Disk systems, tapes, enclosures, drives, libraries, storage of tapes offsite, time spent managing tape and off-sighting the data, power, cooling, salaries, etc… The BIG picture. Compared with the ROI you would receive if you used your current broadband connection, removed ALL management costs, tape costs… then, with the right data protection / replication software, the ability to build virtual copies of the servers in production into a failover target for DR purposes, well you have yourself some intense cost savings. Now that I mentioned DR, or Disaster Recovery, we arrive safely at our last stop on the Top 5 list this time, at Number 1, Business Continuance.
Number 1 – Business Continuance
So the number one reason to replicate data is not surprising. Business Continuance is what we do when things go bad. Every company needs a plan a, a plan b, a plan c. As we have discussed so far, with the right solution, the lines between data protection and replication are converging, as they should. You use both to survive major failures. Combined, theoretically, they are aware of each other and can cross the lines easily.
Back data up from these servers, store it locally in a compressed and deduplicated fashion, send the data via some secure and encrypted pipe in its compressed and deduplicated state to reduce WAN utilization, and store a copy in the remote site. As a bonus, the remote site would contain a few servers that house virtual clones of the production environment that automatically get updates when they arrive from the production site.
This is a reality today. It removes the manual intervention, the human component, and provides the ability to failover to the remote site in minutes with minimal data loss. Data is automatically placed in both sites. It is available at all times.
At the end of the day, a combination of today’s wonderful world of cost reductions, terrorists and natural disasters, and so many other factors, it is time to take a look at the world of replication from a new, fresh perspective. Look towards reducing the costs by focusing on the Total Cost of Ownership of the current environment, not just the cost of a new product or the thought of replication and the way it used to be. Look for products that integrate with your data protection products and if they don’t, time to rethink that one.
What are your thoughts on using Replication as a means of protection? Please share your comments/experiences below.