Arkphire provides IT infrastructure and infrastructure services, advising companies on their IT strategies. Its CTO Howard Roberts shares his view of technology trends and how the IT professional’s role will evolve.
You’ve already moved a lot of Arkphire’s systems off-site to a data centre: has that made you more open to cloud services?
There are certain parts of our application stack that port well to the cloud and parts that don’t. It’s not just always on cost: people are being persuaded to look at cloud-based services too much on cost only and that’s not the only metric.
The main reason for moving our service management platform to the cloud was the flexibility it offered; cost was less the issue. If we’re looking at other systems, like our internal financial system, we didn’t port that because we need local performance for that, but our CRM system is a cloud system and it works perfectly well. So it depends very much on the use case.
What’s your assessment of Irish businesses’ readiness to move to the cloud?
I don’t think all of it is quite ready for the cloud. Most companies’ application stack has not been prepared to be cloud ready. Email ports to the cloud quite easily, where ERP or custom-written applications developed over 10 or 15 years are not so easy to change. It depends with different companies, what applications they have and what stages of evolution they’re at.
Also, there’s the fact that you need to change the whole way your IT people are used. The internal IT staff has to change their delivery model when you move to the cloud: they become service management brokers rather than IT infrastructure specialists which is currently what people are employing. They’re handling service contracts rather than handling physical hardware.
How much of a difference is that to what went before?
It’s almost a completely different person that is needed – it’s a different psyche. There has to be an appetite in the organisation for people to be retrained to a different kind of model as to how they use IT. You’re looking for people to measure service contracts to an external supplier and that takes a different skillset, ordinarily. Most people putting stuff out to the cloud haven’t thought about how they measure the service level compared to the cost and effectiveness of having it delivered internally.
Internally, your expectations of how IT can serve the business might be a particular way: an external party providing same thing – is not exactly like for like. If you’re putting things out of mail services out to cloud, you can measure that pretty simply in terms of uptime and so on, but more complex systems take complex measurements. People tend to take the low-hanging fruit.
What’s happening at the moment is, people are getting an appetite for what they can do. The avalanche hasn’t happened yet but it will, and when it does, the change will be swift.
What implications does that have for IT professionals?
The jobs are still there, but they’re in a different place. The infrastructure jobs will be in data centres; there will still be the same need for the guy with a screwdriver, but it won’t be in individual companies anymore. Now, it will be about the ability to map internal requirements to a cloud provider’s capability. People don’t understand the cloud provider’s abilities well enough yet.
What are the implications for what Arkphire does, in that scenario?
Mapping the right requirement to the right provider is more what we will be doing, going forward. It’s a sea change.
The large-scale build-outs, there’s not as many and I suppose every person looking at new infrastructure has to look at cloud alternatives, so the evaluation process is taking longer.
For any companies in the same realm as we are – integrators – you need to make sure you’re catering for that demand. You’re selling services to people, not infrastructure.
Given this trend of moving computing hardware off-site, will that change the ratio between what organisations spend on maintaining systems and what goes on innovation?
It’s an opportunity to rethink but at moment what we’re seeing in the financial crisis, most companies have been having a 90:10 split. That doesn’t allow a company to take real advantage of IT. You need to be at a 60:40 or 50:50 split where innovation has room to breathe.
Since 2007/2008, most people are in a situation where they’re just keeping the lights on. We’re starting to see more activity around people buying around need for competitive advantage. We’re seeing that coming back in the last 18 months.
What’s your view of some of the main technology trends right now, such as cloud computing, big data or ‘bring your own device’?
There’s genuine merit in them, but again, it’s about use cases. The ‘big data’ thing has caused confusion, because it’s thought of in terms of the amounts of data rather than the value of the data you’ve got. At the moment, the investment required to extract value doesn’t cover the cost. Until that comes down, for companies our size in Ireland, it might not be an option …
A lot of companies are starting to produce analytics tools that are more affordable, and that’s quite a change. But again, it needs education for people to understand what they can get out of these systems. When we’re looking at Hadoop, for example, tools like that are crunch engines. The real goldmine is in understanding what you’re trying to look for.
You’ve seen a lot of technology trends come and go – based on what you’ve seen up close, what technology has been the most useful for businesses, one that maybe doesn’t get the hype or attention that others get?
The one thing they should have caught a lot earlier was desktop virtualisation, and it’s probably mostly due to primary licence restrictions. Tactically, it’s all there, but financially it’s not been as attractive as it needs to be.
A lot of the traditional software licences don’t allow the technology to advance at the rate it should. There’s very few workloads that can’t be consolidated, but licence restrictions don’t allow you to move to the virtual world. The licensing model doesn’t work so well, so you get workarounds. The true benefit needs to come. Like the music industry that didn’t see digital coming; if they had had a different mode, they would have been on board with the change.
That’s a frustration for us. The technology is simple to do and it’s easy to deploy, but the costs don’t add up yet. A very robust technology can’t really be used in anger yet, except in the larger type of sites.
We’ve recently seen how the US ‘Obamacare’ website launch went badly. Why is it that organisations still get IT projects so wrong?
I guess it’s probably back still to the traditional things: technology companies listen to what they want to hear instead of listening to what companies really want. It’s about translating a business requirement to a functional requirement.
When we talk to companies, and there’s a competitive situation, we regularly see there’s parts of the conversation that have been carefully omitted from the discussions to make the cost look better than it actually is. If I was on their side of the fence that would really frustrate me.
For IT leaders that are extremely budget-constrained, what are some smart and cost-effective ways they can deliver more with their IT without necessarily spending loads of money?
When we were getting into outsourcing and managed services, we had to adopt standards to measure against; adopting ITIL was fundamental for us, to make certain we were looking at a process of continual improvement. You’ve got to constantly look at how you can better the service offering to the customer: it’s not just about securing the deal.
All of our people need to be ITIL-certified to understand the method of how we operate. It takes that, to bring in a different view of how the company operates. You need to have good people with good processes. We were at the point where, four or five years back, we could survive without the processes. Now, you need not just to supply, it’s to supply with trust and customers need to understand your process.
And the key to that is having a benchmark to measure against?
If you can’t measure it, then you don’t know why you’re doing it and you shouldn’t be doing it at all. Whatever you’re doing should have a measurable output and a measurable return.