10 Laws of Information Technology - Never Lose Data

1.       Never lose data

2.       Centralize when you can, distribute when you have to

3.       Standardize when you can, customize when you have to

4.       Simple is better than complex

5.       Spend as little as possible to solve problem or achieve a goal

6.       100% utilization of IT assets always yield the lowest cost

7.       Always have a “plan B”

8.       If it CAN break, it WILL break

9.       Volume Changes Everything

10.   Do Not Be Married To Your Technology

I’m sure there are more “laws” for proper management and use of Information Technology, but these are the ones I seem to go back to year after year.  By abiding by these laws, your technology runs smoothly, your business is trucking along, and your IT staff has a life.  I’ll elaborate on each law in separate blog posts.

Every time I walk into a business either as a consultant or an employee, the first this I look at are the data protection strategies and procedures.  In my opinion, this is the #1 law for all enterprises—the importance of the rest of the laws will depend on the priorities of the enterprise.  Data is the only thing that is impossible (or very difficult and costly) to replace.  If any other piece of technology fails, I can just go out and buy another one.  You may know what it’s like to have your laptop hard drive crash without backups (those baby pictures are gone!), but it’s quite another matter if a CIO loses a day’s labor for 300 people.  Clients and staff are not too happy.

So what are some good data protection strategies?  First let me define two terms that I’ll be throwing around. 

Replication – You have two data stores and you copy data from one store to the other.  This can be done continuously or at set intervals throughout the day.  Continuously is better but it costs more.  The advantage of this strategy is that if you lose the primary data store, you can work on the secondary data store without having to wait a long period of time to restore your data.  Staff utilization remains high.  The disadvantage of this strategy is that deletions and corrupted files will also get replicated.  So if you need to restore a file pre-deletion or pre-corruption, you have to also employ the next strategy.

Backup – A backup is not just a copy of your data, it also allows you to “back up” in time to a point where the file was not deleted or was not corrupted.  Depending on your business need and available disk (or tape) available, you can have copies of your file at daily intervals going back six months or more. The advantage of this strategy is that you can retrieve a file that was deleted one day or 2 months ago.  You can also go back in time to when the file wasn’t corrupted.  The disadvantage of this strategy is that it takes time to engage IT and restore the file.  Depending on the volume of data, this could take awhile.

Best practices call for employing a combination of both strategies.

Laptop Data Protection Strategies

1.       External USB drive – you can use a backup program like Windows back up or just copy files periodically to the external hard drive.  I don’t like to use backup programs since you have to restore them to use them.  I’d rather just load files off the external drive using a program like FolderClone.

2.       Online backup services – I really like Carbonite for my online backups.  $55/yr for your entire “C” drive.  This is a set-it-and-forget it application that backs up your stuff in the background whenever you have an internet connection.  My oldest daughter is at college and I would work on her laptop when she came home.  She NEVER backed up her school work to her USB drive, so I put Carbonite on her laptop so she doesn’t have to remember to do anything.  I avoided that crying phone call after she lost that paper she had been writing that was due tomorrow.

Server/SAN/NAS protection Strategies

1.       Whether your data store is an internal server, a storage area network, or a networked attached storage device, the strategies are similar.  Though there are a bazillion different products to execute these strategies.

2.       Hardware – EMC has a very good product and I would say their backup and replication strategies are very close to hardware solutions.  You buy EMC software, but it operates at a very low-level below the operating system.  Good solutions, particularly if you have a large enterprise with corresponding IT budgets.  I also like Enhance Technologies solutions as they are great for smaller budgets.

3.       Host-based replication and backups.  At Little we use host-based backup and replication.  By host-based, I mean that a Windows Operating system is running the applications to perform the data protection strategies.  Double-take is a great product for replicating data.  Tivoli is an equally great product for backups.  By great, I mean they work really well in our environment and they are very cost-effective.  And they are also virtualized in Vmware.

4.       Tape vs Disk – The next question is where do I put all this backup/replication data?  We used to put it on tape but that method is time-consuming to create the backups and restore the data.  Our backup windows (the time it takes to backup everything) where getting so large, that we couldn’t get it done at night.  Disks allowed faster backups/restores and the costs of disk have really come down.

5.       De-duplication technology – these are interesting and I am watching this.  My current assessment is that it costs about the same as just buying more disks.  Rather than putting my data into a proprietary compressed format that only that vendor can decrypt, I’d rather have uncompressed files that any vendor product can read.

6.       The main thing to remember is that backing up is the easy part—finding your file and restoring it is the hard part.  Make sure you regularly test your restore prodecures.

Geographic Separation of your Data

1.       Once you have your data backed up and replicated, it’s also a good idea to separate your primary and your secondary data store geographically.  The only problem with this is that it could cost a lot of money for a telecommunications circuit to connect the two stores.

2.       It is one thing to have a copy of your data in another place for a catastrophic outage (like a fire in your server room), but it’s another thing to have an online, readily accessible data store that is immediately available for a business continuity scenario.  If you lose a windows volume, you want the business to continue immediately.  If you have a slow connection to your secondary volume, it will make working on it difficult or impossible.

3.       At Little, we brought our secondary data store back from the data center because we only had a 10mb data connection and couldn’t afford a faster connection.  If we had a gigabit connection, the two data stores were in sync within 5 mins.  With a  10mb connection, the stores were at least 8 hours out of sync.  We still have separation of the data stores by placing them in two separate buildings on two separate floors on the same campus.  Then we connected with them gigabit fiber in order to get the throughput.   One day, we will be able to afford a gig connection that is a long distance away (did you see that Google announcement for gig internet service?) and I’ll be able to move my primary/secondary volumes further away (to another city).

These are the best practices for data protection and can be adopted by all sizes of firms and budgets.