Thanks for all the votes and comments, it’s obvious that some of you have strong opinions on this, so I should explain that I deliberately worded the question in the way that I did because I wanted to provoke a reaction.
Here’s how the voting went…
87% of you think that Data is the new Utility
13% of you don’t
And my position? well I’m with the no votes, I don’t believe you can class Data as a utility, ‘Steve’ actually picked up on this in his comment, ‘Data isn’t fungible’ (thanks for introducing me to a great new word). With all of the other utilities Gas, Electricity and Water their elements are fundamentally the same, a litre of water is always a litre of water, and it doesn’t matter what order the elements are delivered in, but with Data although it is just a stream of zeros and ones its how these relate to each other that determines whether you receive your email, Word document etc. 1Gb of data is not the same as every other 1Gb of data
But the point of this post is that I think there are many lessons we can learn from these other utilities, there are some important common considerations…
Cost – We always look at the cost of having utilities delivered to us and as a society we consume more and more of them. With data continuing to grow at exponential rates it’s absolutely critical for us to work out how to control the costs associated to this, the Storage Infrastructures that we build have to deliver higher levels of efficiency and other related capabilities to reduce costs. Being able to tap into the resources of Service Providers and the Hyperscale providers such as Amazon has to be one of our considerations.
Availability – in the same way that when we flick a light switch we expect the lights to come on, or when we turn the dial on our cookers we expect the gas to flow, regardless of whether its a mobile device or an application we expect Data to be there, 24 hours a day 7 days a week. This is pushing companies to consider new ways of building infrastructures, and there are lessons we can learn from utility delivery.
Service – Every so often something will go wrong, its much less frequent today than it ever used to be, its typically quite rare for us to experience blackouts or interruptions to our water or gas supply, but when something does happen we want to know that the company that’s supplying us has the people and scale to be able to deal with it fast. Service is difficult to measure because its often not something that we put a value on until something goes wrong, at that point we find out how good or bad the service actually is and realise very quickly the cost impact.
New Opportunities – (I’ll post a vBlog on this in a couple of weeks). As well as doing all of the things above really well, utility companies are at the forefront of exploring new opportunities, new ways of generating revenue and profit. When the utility that you deliver is just a commodity and not differentiated from your competitors its absolutely critical that you find new ways to position your value.
Nowadays for many things, IT departments are the same as service providers, so think about it, if the service you deliver to your customers is email then what’s your differentiator? water is water, gas is gas, electricity is electricity and email is…well email, you will have to compete on cost and at some point someone bigger will always do it cheaper. Yes, I’m sure you’ll deliver great availability, but so will everyone else, and maybe your service is just great, but when was the last time YOU contacted your helpdesk regarding an email problem? These things will be expected, it’ll be what else you can do, what New Opportunities have you found that can deliver value over and above the service that will set you apart and this is exactly what the utility companies are working on.
Great post Matt. I think that Cloud Providers will have to get good at explaining why they are different than other Clouds in a hurry. Much like in the early days we at NetApp explained why companies should use NAS/SAN instead of DAS…then had to shift to why NetApp for NAS/SAN after SAN became a given. Then Virtualization Vendors were focused on why use VMs rather than the old proven Physical Server model, then had to quickly adapt to explain why use their hypervisor or all the other hypervisors once virtualization mainstreamed. Now it isn’t about Cloud or not. Going to a cloud is becoming a given. Why should I go to your cloud? What makes your ITaaS different? There is still lots of room for providers to differentiate themselves.
A utility is defined as a service, not what is being delivered. Infrastructure resilience needs to be the focus. Applications and users want “internet tone”, but who can afford to build a dial tone infrastructure? I don’t know how many 9’s is dial tone, but it is a heck of a lot more than 5 or 6. Ubiquitous computing has created the perception of internet tone = dial tone. Who is providing isn’t the question when it comes to availability, no one wants to accept less to get it somewhere else. Do you see cDOT is an inflection point on the path to internet tone?
F1 – just great to have it back. Old sound was better, we’ll get used to the new one. Fun to watch those guys try and control all the new power those cars are putting down.
Steve, I’ve been thinking a lot about your comment for the last few days before posting a reply.
Dial tone resilience for data or services is going to be a long term challenge and I think the solutions that are being built today are steps toward this and not the complete solution. For example if we consider Google, Amazon or even Microsoft as examples of Data / Application utility companies, then as we’ve seen very publicly, they still fail, albeit much less frequently than ever before. I’ve not done and won’t be doing any detailed analysis but, are they as reliable for what they deliver as the dial tone electricity / gas / water companies were 10 years ago? 5 years ago? or even today?
I do think that cDOT represents today, a way to build a storage infrastructure that minimises the impact caused by hardware failures (10% of the reason we lose the dial tone), but it’s also a huge step toward removing the impact of the other 90% of failures caused by planned outages (upgrades, hardware or software, hardware lifecycle refreshes etc). If cDOT could provide a platform where a workload could be offloaded to a service provider to enable even more major changes then yes I think you could see it as a pretty significant step forward toward a ‘dial tone’ level of service. It’s certainly more than I see from anyone else in this space right now.
I miss the screaming V10’s and V8’s in Formula 1, I’m a car nut and the noise was just such a major part of it for me that I’m going to really struggle with the new engines.