By Keith Joseph
“Storage Virtualization – Storage Hypervisor – Software-defined Storage.” It has many names, but has the time come for the data centre to embrace the concept of full vendor-independent storage management?
Over the years we have seen a move by vendors to add new storage products to their portfolio by in some cases obtaining – at great cost – new and latest cutting edge storage products, or by signing OEM agreements with suppliers to fill the gap in the product lineup; and some even designing and building new product lines themselves to give them access to a market sector. A case in point is the move into the Solid-State Drive (SSD) market space and the monies changing hands for vendors or from investors wishing to cash in on this current hot market.
We see this with one ‘tier two’ vendor being offered $150m in VC funding to help drive their business forward – or from a large networking vendor paying $415m for a start-up with limited sales because they wished to move into this space. But the signs are clear that’s where they think the market is going, and they will invest to win. However in this modern age gold rush we are seeing companies adopting the stance that ‘if we build it (or buy it) they will come’… but is this the case? Well, the reply from Storage managers is “yes and no”…confusing – yes! Understandable ? Actually, yes. To understand why its confusing read IT Managers Struggling to See Through all-Flash Storage Myths.
So let me put my point of view forward and explain. The storage market is driven by a few big players who will go to almost any length to protect their market place. The reason is they make a very nice return on investment for their investors and this in turn makes them happy – so entitling them to better rewards. Simple that’s the world we live in today. However every now and then a new company starts up that pricks the ears of their customers or a new technology like SSD or flash comes along, so all the ‘hunters’ circle looking to spear this fish in the proverbial storage barrel, and in most cases it’s a win win: Vendor gets new product line and the start-up’s investors bag a lot of cash. However this “win win” for the investors and vendor is a problem for the end-user and the reason is the vendors increasingly have a mixed bag of products they offer.
If you look at most of the vendors offerings out there this mixed bag is pulled together with different interfaces and or management consoles that do not talk or work with each other and in some cases mismatched functions but look alike. Add to this the fact the vendors like a nice cozy closed ecosystem so don’t want to open up their customers storage because by doing so this may reduce their potential to earn big returns – and could also highlight their short comings in the product range.
However saying all that the world of storage is moving fast and picking up pace, and its being driven by the way we consume more and more data and how we are use it, store it and demand it; but the difference now is that we are not willing to pay through the nose for it. Plus the drive to more outsourced infrastructures and data centres, and the reduction in budgets of “on-premise” hardware. This leads to a need to reduce upfront costs and sweat products hardier throughout their lifespan.
So what to do? Well vendors have noticed and are trying to act without fully throwing open the door to other vendors hence the description ‘software-defined storage’ becoming more common in the marketing stuff they put out. Why this term and not “Storage Virtualization” or “Storage Hypervisor”? Well when you have pushed back on a concept for so long it looks a little odd to do a 180 u-turn so they are moving 90 degrees first! Then time will let them complete the move.
So is this the right move? Well for storage consumers it most certainly is. It covers and sorts out a lot of headaches. It removes the mixed management problem, lets different storage types work together, adds new functions to models that are missing them or to old models that never had them – the list goes on…
Does it help the vendor that embraces the movement? Well that’s down to the vendor; if they embrace it fully then yes as the conversations with the storage consumer changes from a “we sell tin” model to a “how can we help or fix your need?” approach. If they do not then the storage consumer will walk and SDS lets them do so without impact.
So who should be looking at this? Well, almost all data consumers outside of users just needing a basic file store. Why? Because it gives options which you do not have with a tin-based offering. Tin is Tin and Disk is Disk, software is the controlling conductor or your storage orchestra.
…who should be looking at this? Well, almost all data consumers outside of users just needing a basic file store.
Is it the magic bullet for the storage world? It could be but for a couple of bumps to navigate around. Cost: currently vendors charge way too much for a piece of software – in some cases six figures plus. Add to this the way they sell it by making you buy a node/ header or two or three then adding the cost of the storage capacity you need to manage – in some cases this can be 2-3 times the cost of the physical disk …so not a good start. Then who should you buy it from? IBM has the most mature offering from the major vendor stable with the SVC range or you buy from boutique software house that are based on different operating systems ranging from Windows, Linux or ZFS. But the choices are out there and it’s down to you. The interesting fact is this will be on ongoing as vendors position and counter position all in the interest of defending their corner…