While the IC’s research organization looks into adding security to cloud environments, in the here and now, intelligence agencies are sharing more data.
Backups are usually a thorn in the side of most IT shops — especially in small organizations and offices.
Although the government is home to some of the largest organizations in the country — comparable to major corporations and manufacturers — with vast technology enterprises to match, there are also many small agencies, commissions, councils and boards under the federal umbrella. Then there are the satellite offices, which may or may not rely on the storage systems and data centers of their parent agencies. For these government entities’ IT operations, managing backups commands a major chunk of their attention.
Take it from someone who knows, tech shop staffs dread spending the time necessary to create these backups; plus, the cost can be enormous. But with the advent of disk-to-disk solutions, checking the logs every morning and swapping out the tapes will soon be a thing of the past. As storage pricing comes down, it may be time to look at upgrading older backup approaches to something more robust and flexible. Even so, it’s not time to let go of that precious tape system just yet.
Here are a few ways that should help your organization make a successful move toward disk-to-disk-to-tape backup.
Application selection will play a major role in the completeness of your organization’s backups, the time spent managing them and the thoroughness of any data recovery efforts. It’s wise to spend a lot of time evaluating products to find the one that meets your particular needs.
The initial upfront investment makes the value of a little research worthwhile to avoid headaches later on. A few things to look for: the ability to use versioning, a user-friendly interface, thorough reporting and stability.
One way to save money is to have a central server that you can pull all of the data to before the backup software takes over. For most backup systems, application charges are based on the number of servers or workstations, which can quickly add up to big bucks. If you can take your SQL server or a web server and copy the data to your main backup server before you run a backup, then you can save money on purchasing licenses for those servers and on yearly maintenance contracts.
In my shop, we use the utility rsync to grab data from most of the servers and consolidate on a main backup server before the backup app stores files to a tape system and manages versioning.
Most agencies have tape backup systems in place, so continuing to use them will not increase the budget. Sure, you have to pay to keep the warranty current on the hardware and buy tape when needed, but those are minor expenses and probably already built into your budget. Plus, it’s easy to store tape offsite: Tape can be stored just about anywhere, doesn’t take up much space and offers a great secondary backup. You’re not likely to get in trouble for having too many backups, especially if disaster strikes.
The best thing about using your tape backup system as a secondary solution is you can run backups whenever you want because you will be backing up from the central backup server. You will not use any network bandwidth, and you will not have to take servers offline to complete the backup.
When people first consider disk storage, there’s a tendency to assume that an expensive storage area network will be necessary to house the disk-to-disk backups. While a SAN does allow redundancy and boosts throughput, it’s not essential to move to disk storage.
For disk-to-disk backup, there are a few main drivers, and they don’t necessarily require that your agency implement a SAN.
First, you need a lot of storage. Take the entire amount of data you want to back up and multiply it 15 or 20 times: That’s how much capacity you should plan to have for your current backups. You can always plan for more, especially if the amount of data you store has been increasing steadily over the years. But the reason the 15-to-20-times figure works is because we don’t do full backups every single day. By using incremental, differential and versioning backups, you can streamline and store only the changes made to files, not duplicate complete files over and over again.
Second, you need hardware that’s prepared for any type of failure. Using some sort of RAID configuration is a must, and purchasing a system with redundant hardware also makes sense to ensure continuity of operations. For instance, I buy extra hard drives that I keep in case I have one or two that break down in my configuration. RAID 6 is my choice because it provides dual parity and will protect against failure even if you have two drives die in your configuration.
There are many ways to craft a backup schedule, and a lot of outside factors will affect what you choose to do. You should consider the size of your backup window, how it affects available bandwidth, whether the backup will require server downtime and how often certain servers need to be backed up.
Let’s say you have five database servers that need to be backed up once a day only. You can schedule these backups to happen at different times and then move the files to the main backup server. Using the existing tools in most databases, you can back up a database without having to affect users. If you stagger each backup and copy job by 15 minutes, you will prevent an overload of traffic on the network during business hours.
Another thing to think about is how you want to manage your full and incremental versioning. In the past, with tape systems, most people made full backups at the end of each week and each month and incremental backups during weekdays. With versioning and fast disk-to-disk storage, you can back up a lot more data more often. If your backup software can do versioning, then you have flexibility in how long you keep your backups.
For example, if you have backup jobs that run daily, weekly and monthly, you might set the expiration for the daily runs at two weeks, weekly backups at six weeks and monthly jobs at two years. This scenario means each backup overlaps another and protects your data against any unusual or unexpected failure. Plus, with versioning, you can make efficient use of available hard-drive space.
Think of versioning as keeping a running log of all the changes to your files over a length of time. Let’s say you have a file that you work on almost daily. Something happens to the file that caused some ill effects about two weeks ago, and it has continued to deteriorate. With versioning, you can pick a certain point in time from which to restore the file.
Versioning helps make backups more efficient because it allows you to create one full backup, and then back up only changes to files. This leaves you with more available storage, so you can buy less storage capacity up front or keep your existing backups for a longer period of time.