World Backup Day
According to Wikipedia, the term "Hallmark holiday" is used predominantly in the United States to describe a holiday that is perceived to exist primarily for commercial purposes, rather than to commemorate a traditionally significant event.
One of my pet peeves are the made up “special occasions” that greeting card companies and others make up to drive sales of more $5 greeting cards. I’m all for celebrating mothers, letting my wife know I love her dearly, and otherwise acknowledging the important people in my life. But I don’t want a ton of commercialism around it since it cheapens the sentiment.
World Backup Day is one of these “Hallmark” events. You should make sure you have backups of your data and you should probably have a better plan in place you do. But this isn’t something you should only think about on March 31st each year. According to ESG’s 2012 IT Spending Intentions Survey improving backup and recovery is one of the the top overall IT priorities for 2012. The sad reality is that World Backup Day is promoted by backup vendors -- these are the ones that have failed to relieve IT’s worst pain over the last decade or more.
The reality about backup is that it can’t be solved by putting lipstick on the pig that is backup. As an industry we’ve been trying to do that for many years. We have made the pig look better with great technologies like compression, deduplication, disk to disk backups, etc. But in the end we have an entire class of storage that is growing out of control, has value only when its needed, is a pain to manage, etc. It’s time to think out of the box. Instead of making backup better, what if we eliminate backup as we now know it?
At Nasuni, our premise is that the best solution to backup is not doing a backup. Note that we’re not saying you don’t need to protect your data, we’re just saying that backup is the wrong way to do it.
So what is the purpose of the backup function? First, let’s define backup:
A backup or the process of backing up is making copies of data which may be used to restore the original after a data loss event
For unstructured data (80% of the world's data today) this amounts to protecting from all the various forms of data loss which includes things like accidental modification/deletion of files, intentional malicious changes, crashes of software/hardware, etc.
To protect against all these forms of data loss and solve the need that has traditionally been defined as backup you need to do the following:
- Maintain copies or versions of the data that you can recover after a data loss event.
- Provide an interface to create the copies and access the copies (for restore, litigation support, etc.
- Keep the copies in a form that cannot be modified -- you can’t protect against tampering if your backups can be tampered with.
- Keep the copies in a secure location that is offsite to protect against some of the data loss forms.
In recent years we’ve seen some of the traditional storage vendors come close to solving the backup problem by making the move to eliminate backup. Their approach has been to use snapshots and then a replication scheme between multiple sites to solve the offsite need. This is a costly approach to solving the problem and unfortunately it has several flaws. Snapshots in the traditional world consume resources and take away capacity from primary storage. These snapshots are limited in quantity (perhaps 255 snapshots if you’ve got the space, etc.) and need to be aggressively managed with some kind of snapshot aging strategy. The way the snapshots are implemented to make them as efficient as possible creates a lot of links between the snapshots which means if a file is corrupted on disk in one snapshot (after all it's the same medium) it’s possible multiple snapshots would be affected.
So what do we need?
- Unlimited snapshots
- Snapshots that don’t consume primary storage space
- Snapshots that are stored on a different medium than where primary data is stored
- Snapshots that are protected with modification detection and correction algorithms
Storage services offer unlimited snapshots of your data. These snapshots are generated and controlled by the Nasuni Filer but stored in the cloud component, which is a completely separate storage medium existing offsite. The cloud storage vendors make multiple copies of the data and have modification detection and correction algorithms and so does the Nasuni Filer. Snapshots or versions of individual objects are available on demand thanks to advanced caching algorithms and with Nasuni’s simple user interface or built-in functions like Windows Previous Versions there are simple interfaces to create or access versions of your unstructured data. And to get all these benefits all you have to do is use the Nasuni Filer as primary storage -- that’s right, there is no “backup” operation for IT to perform.
Forget World Backup Day. Backup is no longer an unsolved problem.