Question

I was recently tasked with coming up with an offsite backup strategy. We have about 2TB of data that would need to be backed up so our needs are a little out of the norm.

I looked into Iron Mountain and they wanted $12,000 a month!

Does anyone have any suggestions on how best to handle backing up this much data on a budget (like a tenth of Iron Mountain)? How do other companies afford to do this?

Thanks!

UPDATE :: UPDATE

Ironically enough, I just had the sort of devastating failure we're all talking about. I had my BES server fail and than 2 days later 2 drives in my Exchange server's RAID5 died (2!!!??!). I'm currently in the process of rebuilding my network and the backup integrity is an definitely an issue.

At least now my bosses are paying attention :)

Was it helpful?

Solution

You can buy external eSATA RAID boxes in the 8TB capacity range for $2600. I'm not saying that particular product is the right choice, but that's the kind of box that will do 6TB in RAID5 and still be portable enough to buy a couple of them and rotate them through the bank, like Stu says.

Obviously if you have to have to keep 7 individual days worth, a 14 day, 30 and 90 day snapshot, etc. then things are going to be much more expensive, but it's certainly doable if what you're after is just disaster recovery.

The biggest thing to make sure is part of your plan is actually testing the restoration from the backup. That seems to get overlooked WAY too often and turns out to be the weakest link in nearly all of the strategies.

You should plan for scheduled restorations as often as is reasonable where you actually dump the real data and restore from the backup. Without that, you don't know that it will work when you NEED it too.

I've lost track of the number of times I've been in a company where there's a big rack full of backup tapes/drives, all dutifully made according to the schedule only to find out that NONE of them have valid data when the server gets wiped out.

The more ways you can verify the integrity of the backups the better, but nothing substitutes for doing an actual dump/load from one of your backups to really test the setup.

OTHER TIPS

Amazon S3 might fit your budget better. I don't know if there is software available to automate the backup process but it's rather easy to write your own code to handle this. Here's their pricing calculator.

According to my estimates you're going to be well under the $1000/mo range.

You really have to assess the true value of your data. If you lost it tomorrow what impact would it have on your business? We use offsite backups, it isn't cheap, but if we were to lose our data the business would cease to trade withing 2-3 days.

We considered on-site backups as a possible cost saver but in my experience with data centres/computer rooms over the last ten years (as both an employee and a customer) I've seen fires, fire suppression system malfunctions (wet), hardware theft and one day a car crashed through an external wall right into the suite. Add to that our last DC was located at Heathrow, right next to the runways....you never know what strange things can happen (remember the BA 777 that got caught short of the runway on landing?).

My advice, assess the value of the data then decide if $12k is too rich to keep it safe.

2TB is chump change nowadays.

Look into hard-drive based hot-swappable backup machines, and rent a box at your local bank: http://www.high-rely.com/ (there are many more products such as this, but my Google-time is limited).

Jungle Disk is one such piece of software that can automate the backup process to Amazon S3. I use it for backup at home, but I guess it could work just as well from a server. Also, there are probably other backup tools that make use of S3 for offsite storage.

We've been using DataDomain appliances for that purpose for about 2 years. They're not inexpensive, but compared to $12,000/month they'd pay for themselves pretty quickly.

Basically, we send our backups over NFS and CIFS to one DataDomain appliance, it deduplicates the data and then replicates the differences to the other appliance we have at a remote site.

As for pure online solutions, make sure you do some back-of-the-envelope calculations first. For example, if you have 2TB of churn a month, you are going to saturate a 1Mb Internet connection just for your backup traffic!

As previously mentioned, Amazon S3 is definitely an option, but it may be cheaper in the long run to own the hardware you are backing up to.
For example: Buy a basic server and and eSATA RAID5 setup with 2-3 times the capacity you currently need, then install it at a co-location center. Preferably one with high, but cheap, bandwidth.

This way the server and storage is off-site, but after the initial cost of the hardware, you are only paying for bandwidth.

Granted, the downside to this is that, unlike something like S3, if the hardware goes down you have to go fix it yourself, or pay the CoLo people to. But this may be a tradeoff you are willing to make.

Also, with this solution, you are still going to need a beefy upload pipe to handle the traffic... so there's always the "sneakernet" solution.

I've used bqbackup.com for 1-2 years no problem. You can do a sync using rsync nightly. Wanted to add that their prices are dirt cheap, and I now have close to 1TB with them.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top