The Power of 3
The power of three: Software Defined Object, File, and Block on the same system.
Choosing a storage approach for the modern data centre involves quite a few calculations and considerations. If you’re a typical enterprise, you’ve probably virtualised just about every server you can, and that means choosing a solution that will fit in smartly with VMware and HyperV.
You will already have back-up and recovery requirements and data in different formats in different geographies. You will have tiered data, and probably a ‘cold’ store for the stuff that only gets accessed once in a blue moon, but nevertheless the law or company policy says you have to keep. You might well have a hybrid arrangement with some storage in the cloud, and some for critical applications close to home on high performance appliances. You’re most likely going to have both structured and unstructured data, and, as time goes by, more and more and more of it. And, despite the average price of storage coming down, the average spend on it is still going up.
A few years ago, you would have used File or Block storage for everything. File storage is great for the stuff you keep on the LAN – you could think of it almost as just having a shared hard-drive on another computer for old school collaboration and sharing: the kind of data that sits in xlsx spreadsheets of your company P&L and overhead accounts, jointly worked on by a ton of people in finance and business management; everything neatly divided into hierarchical folders with easy sharing and access control. Block is the workhorse of application storage – you’ll find block in pretty much every single enterprise the world over where there are applications back-ended with SQL databases. . . . which is tantamount to everything. Block based storage is mature, and has a larger installed based than File. Block is the leading approach when performance is critical.
Then of course there’s the new kid on the block, Object. Object storage was originally developed by Web 2.0 companies who were the first organisations in the world to really suffer the pain of exponential data growth. They needed an approach which – like opens source software defined storage from SUSE – could be run on commodity hardware to make some very necessary cost savings. Even if you aren’t using Object in your enterprise, if you’re online at home you are benefiting from it: everyone in the whole world who is online is accessing an Object storage – that’s how you are streaming those 30m+ songs on Spotify, and accessing the gazillions of photos stored by Facebook. It’s amazing for unstructured data (like songs, images and video) and the capacity to have huge amounts of metadata with the data itself allows all kinds of management automation: you can control when data moves from tier one to tier two for example, or even when its deleted. And Object can be distributed across enormous geographies, and theoretically scales infinitely.
So, let’s sum up: the modern enterprise has one hell of a lot of storage platforms using different protocols in different locations on-premise, in the cloud and hybrid. The volume of data is growing constantly, it’s expensive, and there is always pressure to cut costs. That’s why it makes sense to move to software defined, its why – like the web 2.0 companies before you – it makes sense to choose open source, and its why SUSE Enterprise Storage 3.0 supports File (early access), Block and Object – so you can make cost savings on whatever platform or protocol you choose. I think you’ll find its worth investigating.