OPEX – or Operational Expenditure – is the widely accepted catch-all category for all a business’ costs in turning inventory into throughput. This puts it at the very top of the tree of distinctions, up there alongside “animal, vegetable or mineral” - the kind of grand philosophical distinction which is used all too often in business to explain why an issue isn't your problem.
In thinking about OPEX and the Cloud, there's an even grander distinction to draw between those who are fantasising about cloud computing’s capabilities, and those who are nursing burned fingers after an early or mis-intended foray into the topic. First, let's work through the assumptions people have made, and then lay out some of the cold hard facts.
Assumption number one, to beat all others, is that cloud computing trades off OPEX against CAPEX – that is, capital expenditure. In other words, cloud computing means you can forget about buying any more hardware. Just pay someone a nicely appropriate subscription-based fee, and the savings don't come from OPEX at all – it's nil CAPEX, with Amazon or Google buying all the kit on your behalf. You can be a humble tenant, gaining all the economies of scale of a global player's purchasing and architecture.
This may become possible, and somewhat later, even advisable. There's certainly a massive group of salesmen, advisors and even board members who seem to think that time has already come – however, attaining this situation is one thing; sustaining it is an entirely different matter. What starts out looking like a good way to avoid generating overheads and staff commitment can turn into a balance sheet hemorrhage when it comes to public cloud counterparty operations.
Suddenly you’re 100 percent dependent on one supplier, to whom you matter hardly at all. Your needs can diverge from theirs, slowly, and their performance can vary wildly from promised levels, because it's now become so difficult to disentangle yourself that you can't even apply the old supplier tourniquet of non-payment. And that's before you realise you’re now subject to the Salesman's Delight that is “upsell”.
The cold hard fact is that in the short to medium term, reduction in OPEX using cloud computing comes almost entirely from Private Cloud technologies. Old servers eat power – three times over, in fact, because they spit out heat and need to be cooled. Moving a pool of ten servers to four virtual machine hosts and a storage array prepares your internal systems for full cloud computing portability, without making you a hostage to a sales-driven supplier which is finding its way just as much as you are.
Virtualisation is the prerequisite of cloud computing preparation, and such is the pace of change in technology following the end of the recession that firms surviving the impact of the credit crunch are now well positioned to make CAPEX-driven decisions: investing your capital in hardware and software that deploys cloud computing internally first is an excellent strategy for avoiding the marketplace's first-try offerings of SaaS (Software As A Service) while giving your workforce skills in cloud-related topics.
There's another frequently heard hope, especially at the very largest scales: that cloud computing paves the way to unheard-of savings in the field of ICT staff. Lots of operating costs are sunk into smart, expensive people who do incomprehensible things with your network, pausing only to come to you and ask to hire more smart, expensive people.
This is where your appreciation of the vital “human resources” that make up your business can make cost savings seem like an ever-receding target: on the one hand, IT has become a vital pillar of the workflow; on the other hand, only the people inside the IT team have the oversight to know how it all fits together. Cloud technology adoption – whether it's driving an SaaS rollout or a server room refresh – also redefines many of the roles and expectations within your business’ IT department.
You may be able to take out the bottom end unskilled worker costs when it comes to running backup tapes, doing space management and the like – but the middle-tier process experts, the MIS report writers and in-house customisers still carry right on working.
Getting to an OPEX reduction in the first year of a cloud computing project is something of a faint hope; many parts of the process take time to have their full effect, needing business structural change to follow from technical changes to infrastructure, supplier relationships and skill level requirements.
On a three-year view, however, cloud platforms can show appreciable savings: case studies from early-adopters within the computing business (AMD & Yahoo Germany) indicate that a 75 percent fall in OPEX versus a 2005-era platform is realistic.
Steve Cassidy is a consultant specialising in re-engineering networks at all scales of businesses. His affinity for crisis management can be traced back to his career path through N M Rothschild & Sons, where he gained an understanding how people use technology,. Despite the temptation to view subsequent challenges as an anticlimax, Steve has been writing about managing and design networks for the last 18 years. His mix of experience in market economics, project finance, network and telecommunications technologies has informed decision makers in firms of a wide range of sizes and sectors.
