Open Source vs proprietary SDS – why get locked in when you can stay open and in control?

Once upon a time, enterprise storage was defined by hardware. You installed a system, and when it ran out of space or its performance started to creak, you bolted more capacity onto the side or brought in a new, larger system and managed the painful upgrade process, data migration, or both.

You chose from different vendors for different requirements. So, over time, you added additional complexity: RAID, NAS, SAN, File, Block, Object. Different processes and formats for back up and disaster recovery, for the long term ‘cold store’ and the ‘hot’ data you are working on right now. You built a whole bunch of silos, many if not most of which are managed separately – making for a hefty resource issue: management, capacity planning, provisioning, complex error-prone processes, growing volumes. . . . the necessary-yet-dull stuff of IT operational management that will never raise the pulse rate of the Board but whose costs may set a vein throbbing on the forehead of the CFO. Where’s the ROI or the business advantage?!

There are of course (traditionally at least) three headline costs in running storage:

1. Hardware
2. Software
3. People

Today you can add in managed services costs from cloud providers – along with the endless verbiage about OpEx vs CapEx and concentrating on ‘running the business’. Let’s put that aside for a moment, as however you look at it, intellectual property concerns, regulatory compliance, and processing requirements are most likely going to keep a big chunk of your most valuable data on premise. This makes for an obvious business case for SDS (Software Defined Storage) -it’s a bit of a no brainer to reduce hardware costs by moving from proprietary hardware to commodity x86 – even if you are choosing from a narrow list of supported vendors – because you are going to eliminate a big chunk of hardware cost (and lower your CFO’s blood pressure!).

Right. So we know we’re going to go software defined. Now for the more challenging bit: proprietary, or open source? Both are going to reduce cost #1, hardware. What about the other two?

Well, in theory at least, #2 is also a no-brainer. Open Source software is free. There is no cost. Zero. Zip. Nil. Nada. No OpEx, and no CapEx. OK, I say ‘in theory’ because the third factor, the one that makes it complicated, is the human element: #3 people. What, asks your CFO, is the point of taking out software cost, only to replace it with a wedgy consultancy and/or wage bill for the skilled resource that can plan, deploy and manage open source software defined? And that, is an entirely fair question that must be answered.

But does open source really add to your people costs? It certainly requires skills you may not have at the moment, but let’s face it, you probably need those skills anyway. How so? Let me explain.

1. You need new the skills in the business anyway. Pretty much every enterprise in the world is embarked on a journey towards digital transformation. How a company presents itself in mobile via apps, and on the web, is a key differentiating factor and a binary win/lose vs the competition. As we move towards a more agile world, a DevOps culture and ‘hybrid’ IT, the glue that holds everything together is software. You need these skills, and they are complementary to open source. Its in the code, stupid!
2. Your data management is almost certainly going to be open source. Organisations need to do data analytics at scale. Chances are, particularly if you are in a major organisation, your data science team is using or exploring Hadoop, with one eye firmly on Spark. They are open source. They have converging roadmaps and development teams with other open source software projects. As time goes by, the operational synergies will be, literally, epic. Get on board now to avoid expensive catch up work later.
3. Your Development teams want open source. Recruiting and retaining top development staff is difficult and expensive. Digital businesses compete aggressively for the right people. Pop downstairs and ask the project and dev teams what toolsets they are using to do the connections between on-premise and cloud systems. You’ll hear a lot of talk about APIs, and the chances are your version control is run on GitHub – they will tell you plainly the future is open source. They want to work in organisations that look like the future, and they want skills that can be taken outside your organisation.
4. People costs don’t scale in the same way as proprietary software costs. The proprietary model charges more the more you use it. Yes, you will have to add in people costs with open source (but as per point #1 you need those skills anyway, and even if you do go proprietary, you won’t eliminate people cost, or contractors, or consultants.

It won’t be right for everyone of course. And there are definitely organisations where the CFO and the board don’t listen to the arguments of the IT team. Here at SUSE, we think those are the organisations that won’t weather the digital transformation storm: any which way you look at it, in the modern world your IT capability is your business advantage.