By Mark Benson
Data Management has always been a challenge for organisations and with the changing landscape of how businesses run their IT, this has never been more accurate. For years, many organisations have had a siloed approach to data management. And by that I mean data is put into buckets - whether that be on premises on traditional architecture or in a private cloud - through to branch offices with their own silos of data. Add into this the advent of public cloud and the utilisation of multiple public clouds, companies have found themselves with pockets of data spread across the whole organisation. As a result, many companies have approached data management in a similar way - with the approach being siloed, rather than an overarching data management strategy.
Organisations are also trying to get real business value from the data they are storing. It was the mathematician Clive Humby back in 2006 that stated “data is the new oil”, and he couldn’t be truer. But very few businesses are realising the advantages of what they store. Instead, data has become a burden to companies with concerns around data growth, reduced budgets, increased compliance and security requirements, and agility.
So, what should businesses do to address these concerns and start to realise the value of what they store?
First things first, the question to ask yourself is do you actually know what your company is storing? And this is not just “how much data in terabytes” and “is it file or block data”. What are you actually storing? Has it been classified? How long have you had the data? When was it last accessed? And who owns the data from a company perspective? For those of you with a strong information lifecycle management this will be relatively simple, but for those who have not implemented such measures you may need to audit the data and then think about putting in place a policy for the organisation.
Now for some of you, you will be thinking “I don’t even know where all my data is!” And that is perfectly normal. So alongside data analysis there may be a requirement for data discovery to understand where data is located and how it is architected. By going through this process, you are now armed with the information to start to plan for the future of data management.
The next logical stage is to start to look at how you are going to manage data going forward - so what classification are you going to apply to the data. What are the retention periods for that data? (If you don’t know, then this is the time to find out!). How will the data be tiered to ensure cost control and simplified management? If you are not utilising cloud, do you plan to and how will cloud play a role in your data management strategy? For many companies their ILM policy will have been written years ago and may not include Cloud Governance, but this now needs to be a consideration with the advantages cloud can bring.
At this stage, it is also best practice to review your business continuity and disaster recovery plans to ensure they involve the new delivery methods outlined and to make sure they are fit for purpose. Finally, what data are you going to purge that does not fit into the new model? Yes, I am talking about deleting data! We have all fallen into the trap of keeping data “just in case we need it”, but most of the data we store is never accessed and is just a cost to the business, so don’t be afraid to get rid of data. It will also help with compliance.
Now we know what we have from a data perspective, how we want to manage the data going forward, and the models we want to utilise. We also know from the business what it requires from a compliance and business continuity perspective, so does the current architecture meet these needs? If the answer is no, then it is time to look at the technologies available to meet your requirements. Is a 3-tier architecture the direction of travel and what is available in the private cloud market? Is the desire to simplify your estate and look at software-defined storage or hyperconverged, as many organisations are doing. Do you have a Public Cloud provider in place and are you using Public Cloud for storage at the moment? If not, what is available to you? By looking at the market you can quickly choose the right model for your future business and start to plan on implementing this in the future.
This will also be the time when questions of how the storage will be purchased will arise. In the past there were two distinct approaches: the first was to work out how much storage you had, what your growth rate was and then how long do you want the storage for (3 years, 5 years, 7 years?). This would lead to an order for 5 years of storage up front even though you didn’t need most of it for years to come and sometimes you never actually hit the growth rates you were expecting. The other approach was to buy what you need today and then add incrementally over time, but this is a risky strategy as you are guaranteeing that you know the business will have budget when you need it, and that you can quickly get storage as and when. Both of these approaches where largely CAPEX purchases and did not take advantage of financing or OPEX models.
The more common approach now in the market is to buy storage on an OPEX model, so you only pay for what you are using. This allows you to grow your storage over time and just pay for what you are using. It also gives burst capabilities as many models include a buffer that you can flex into when needed. The biggest hurdle for many organisations is the move to OPEX as budgets have not traditionally worked that way in the past, but with the advent of cloud, more and more finance departments are now planning budgets with an OPEX model.
Now if you have read this far (congratulations!) and you are sat nodding your head thinking “Yes that’s what we have in place”, fantastic news! You are truly in control of your data and probably getting real business value from it. However, I imagine most of you are saying “Yes we face that challenge” and that is absolutely fine, because many companies are in the same position. The question is: what do you do next? If you truly can’t see the wood for the trees or are confused as to where to start, please feel free to drop me a line, as I would be more than happy to go through everything in a lot more detail. The thing not to do is acknowledge you have a data management challenge and not do anything about it, as it will only get worse!