Logo Open Source Initiative
Image via Wikipedia

Driving home last night, listening to podcasts from the BBC, The Guardian and John Willis, I began to wonder about the ‘inevitability’ of Open Source in the Enterprise. I shared the initial idea on Twitter and had some useful responses overnight;

“Thought; Open & Closed Source coexist well in enterprise. BUT once you deploy Open Source tool can you ever get budget back for Closed ver?”

Conversation was clearly constrained by Twitter’s 140 character limit, and several respondents (incorrectly, but understandably) assumed that I was simply suggesting Open Source Software to be cheaper. It may be, but it doesn’t have to be. And that wasn’t the point.

For the avoidance of doubt, let me first expand upon that brief tweet in order to clarify what I meant. Then we can explore the issues in a little more depth and see whether the premise holds water.

Thankfully, we appear to have moved beyond the ridiculous polarisation of recent years that saw Open Source evangelists square off against Microsoft junkies and other proponents of closed source proprietary solutions. Naive notions that ‘free’ software would not cost anything to deploy and maintain have faded from the conversation, enabling growth in for-profit ventures to support deployment of all that ‘free’ software. In essence, we have reached (or are rapidly approaching?) a healthy balance; a point at which Open and Closed Source solutions co-exist within many Enterprises, with each solution selected on its merits rather than for the flavour if its religious dogma.

So far, so good; and it is against this background that my question should be considered.

The essence of my question, actually, isn’t one of Open versus Closed at all. The essence is about budgets and politics, and the ways in which organisations manage them. A traditional on-premise installation of proprietary software is generally held to be a Capital item in the budget. It’s a single – and sizeable – allocation of funds once every 3-7 years that results in ‘ownership’ of an asset. Even if an Open Source deployment actually ends up costing almost the same over the lifetime of the installation, the vast majority of those costs are in terms of people who are either employed or brought in as contractors from outside. Those people are recurring items of expenditure in the budget; smaller amounts, more often.

In choosing to procure a new piece of software, it probably doesn’t ‘matter’ whether it’s Open or Closed. Instead, the decision should (correctly) be made in terms of the ability of the chosen solution to meet a set of business requirements. Sometimes the best solution will be closed source, sometimes it will be open.

But once an enterprise has opted for an Open Source solution, various internal considerations take over;

  • the budget holder (probably) gains a number of staff dedicated to supporting and customising the new solution, either on the payroll or as external consultants. The size of the budget holder’s empire grows visibly larger;
  • the CFO sees the unpleasant pain of large Capital investments removed from their books, replaced by a more manageable steady trickle of recurring costs each and every year.

At that point, I’d argue, it must effectively be impossible to turn back from an Open Source solution to a future Closed Source replacement. The switching costs, perceived loss of ‘power’ and perceived hit to a bottom line accustomed to the steadiness of recurring costs all combine to dissuade numerous stakeholders from making the switch away from Open Source. Any proprietary competitor would have to be far better than both the incumbent and all the other Open Source alternatives to stand any chance at all, and in a mature software world where most applications are actually pretty comparable that seems unlikely in the extreme.

So, I’d propose, the decision to move to Open Source is actually made on a reasonably level playing field but the decision to move away from Open Source takes place in an environment so stacked against Closed Source that it’s unlikely to happen. If that is true, we will see a gradual movement of various Enterprise applications towards Open Source as purchasers select each application on its merits (inevitably, some proportion of those will be Open Source). However, we’ll see almost no movement the other way, leading to the eventual dominance of Open Source across the Enterprise. QED. Or maybe not?

And, for some reason, my background in Archaeology comes bubbling to the fore, offering the Roman Army’s fossa punica as an analogy; you can get in [to Open Source], but you can’t easily get out.

An email conversation this morning suggests that vendors of Closed Source solutions are aware of the threat, at least subconsciously. Rhys Wilkins, for example, pointed to the relatively low incentive for piecemeal replacement of software from companies such as Microsoft that tend to bundle ‘desirable’ and less desirable products together in packages. Their customers are prepared to continue paying for the desirable products (Exchange, say) and as they have to keep paying for the bundle whether they use all of its pieces or not there is a disincentive to gradually bring in alternative offerings.

Hosted subscription applications (SaaS) such as Salesforce only serve to confuse this picture still further. They’re not Open Source by any means, but in budgetary terms this new generation of subscription software is a recurring rather than Capital item of expenditure. Will we therefore (as Justin Leavesley suggested in an email) effectively see a race between those trying to push Open Source Software into the Enterprise and those trying to evangelise the benefits of entrusting previously internalised processes and applications to proprietary Clouds? And what happens when Open Source meets Open Data and the Cloud, and some of today’s proprietary silos begin to leak out onto the open Web?

But that, I think, is probably the subject of a subsequent post.

Reblog this post [with Zemanta]