Not so long ago, hardware-centric solutions seemed to be the only answer to storage-related issues. If additional storage was needed or performance had to be improved, there was only one option – throw more hardware at the problem and hope it helped. In most cases, it worked, but in general it created vendor lock and silos of storage that needed special skills to manage. Is solving today’s storage problems by hoarding hardware solutions still the best approach in the modern world?
Previously, the server market was purely hardware driven with applications sitting on expensive, discrete physical servers in a scale-up architecture. But the emergence of server virtualisation and scale-out technologies has given traditional storage vendors a real problem. Storage vendors may acknowledge new technologies and even market them, but their revenue streams are mostly built on legacy traditional storage arrays.
As we are seeing in today’s declining hardware-centric environment, the industry’s move away from legacy hardware arrays and the support and services that go with them are pushing their revenues to a consistent decline. Ultimately it is not in their interests to allow existing customers to move away from traditional branded storage.
Protecting the bottom line
The truth is that the pressure of delivering shareholders’ profits and meeting revenue commitments is dictating the product strategy of traditional storage vendors. New storage technologies like software-defined storage (SDS) offer flexibility and scalability at lower pricing on commodity components. However, advocating a completely hardware-agnostic SDS option to their existing customers would result in lower revenue and profits for traditional storage vendors. They have financial incentives to keep customers locked in to traditional storage for as long as possible or provide a SDS solutions with a new licensing models that favor their own underlying hardware.
Time to break free?
The challenge with blindly throwing more hardware at a variety of IT problems limits organisations flexibility and leaves the customer with the only option of buying even more of the same solution. In terms of storage, IT teams are led to believe that if they want their data to move freely around on internal servers, it will need to be on the same branded hardware that they already have. However, the right software can offer other alternative solutions to this problem. An horizontal SDS layer hands power back to the organisations by sitting above various hardware resources and managing it all, no matter what brand it is.
Companies that have implemented massive scale-out data centres built on white box hardware solutions have quickly embraced the SDS model as it gives them scalability at a low cost, along with other benefits such as lower management and maintenance overheads. It is common to see this approach in large organisations where the amount of data being stored is high and likely to involve numerous disparate storage platforms.
Change of focus
As SDS platforms mature and provide more robust data services, it is getting harder to justify just buying more of the same from a hardware vendor. Organisations that take the plunge and opt for a software-first approach will benefit from greater choices when making their next purchasing decision as well as the freedom to pick the most suitable product, rather than be restricted to what is compatible.
As well as improving the bottom line of an organisation, software can be an intelligent layer that brings other benefits, such as data recovery, migration and optimisation of the data being stored. When organisations see the benefits and low costs of SDS, it will be hard for traditional vendors to justify their legacy hardware. A fully automated, remotely managed data centre that drives appropriate total cost of ownership (TCO) targets for profitability, can only be achieved through software management and hardware abstraction at every layer of the infrastructure from servers, to network to storage. Converting to a software abstracted storage ecosystem is the only way to truly drive TCO, lowering costs and providing visibility into performance and usage statistics of the storage transactions throughout the data centre.
Author – Farid Yavari