Has open source software’s day finally arrived?
Dr Graham Oakes looks at the benefits and barriers of open source software adoption in UK public sector organisations
The Conservatives came into government with a manifesto commitment to “create a level playing field for open source.” Before that, Labour published its ‘2009 Action Plan on Open Source, Open Standards and Re-Use.’ With an imperative to cut costs and cross-party support, open source must be flooding across Whitehall, mustn’t it? So where do we stand? Here are some of the realities behind open source.
The cheap option?
Say open source and the first thing many people think is no licence fees. The attractions are obvious: licence fees are a highly visible element of many software projects. Eliminate them and you make some significant cost savings.
Sadly, it’s not that simple. Licence fees are the tip of the software iceberg. Installation and configuration can add significantly to overall costs. Training adds yet more. And the organisational change management necessary to actually deliver benefits can dwarf all these. If open source inflates such costs, it may not be cheap at all.
In reality, open source can create costs in different areas. The range of products itself creates costs. In web content management, for example, there are a dozen credible open source products. And comparing them requires work: you can’t send out an RFI, as most of them don’t have a pre-sales organisation to respond. You have to do the analysis yourself. (The cost of pre-sales is built into proprietary licence fees. You pay for it whether you need it or not. Open source just unbundles this cost, making it visible.)
Pilots and proofs of concept may also create costs. These represent good practice when selecting any software, proprietary or open source, but you may need more of them for open source: the infrastructure of documentation, training companies, etc, tends to be less well developed.
There may also be costs involved with configuration and integration. Open source doesn’t mean non-commercial. The vendors still need to pay their mortgages and they often do this by charging for configuration and integration.
Customisation may also cost. When developers have access to source code, there’s a risk they’ll tweak it more than strictly necessary. (This can also make it more difficult to apply upgrades in future.)
Support may also be another area of cost. Community support for some open source products is excellent. Even so, many organisations want the certainty that comes with service level agreements. This requires paid-for support, another area where open source vendors make their money.
When you explore such costs, you may find that open source can be pretty expensive. It may still be cheaper than proprietary software. Or it may not. You need to do the analysis.
It may be Low Risk Software
As you do this analysis you may find that open source has a very different spend profile to proprietary software. Licence fees are often front-loaded: there’s a large initial sum then a smaller annual support fee. Open source incurs some upfront costs for product selection, but it never has that lump sum. Instead, you tend to proceed incrementally. After each increment, you decide whether to commit further.
Phasing commitment reduces risk. Consider what happens when you commit to a large upfront investment. In order to justify the investment, you look for additional benefits. Each of those requires changes to the system. Scope creeps. Before you know it, you have a much larger project. Yet there’s one thing we know for sure about software projects: large ones are much more likely to fail. If open source helps avoid this trap, it can save a lot of money.
Open source may provide other benefits too, for example, through reduced lock in. Open source is distinct to open standards, but most open source products work naturally with open standards. This can reduce the cost of integrating with other systems. It can also reduce end-of-lifecycle costs to move from one system to another.
Other benefits include easier integration. Encapsulation and information hiding are generally signs of well-designed systems, but sometimes you need to see how something works in order to understand how to work with it. Open source facilitates this.
Fewer licensing issues is another benefit. Proprietary licensing can require complex usage tracking and accounting. Open source rarely requires this. This is a growing issue as organisations move to the cloud: tracking licence usage as the number of virtual servers expands and contracts in response to demand can be a major constraint to deployment of proprietary applications.
Support for community-based innovation is another benefit. Open source often works well for web 2.0 applications. It can also work well when developing systems that will be shared between non-competing organisations: open source licensing facilitates management of common intellectual property.
Most of these benefits come from reducing risk rather than directly reducing costs. But reducing risk is often the best way to reduce long term costs.
A Growing Range of Applications
Open source has traditionally been strong for techie tools. A programmer writes some code to solve a problem he’s experiencing. He shares it with his mates, and an open source product is born. This model expanded to cover technical infrastructure (Linux, Apache), and then moved up the stack to databases and so on. Highly mature open source products are now available for many technical applications.
Products are now emerging for business applications too:
• Web content management. Open source grew in parallel with the web, so open source content management systems (CMS) are often as mature and functional as proprietary ones.
• Document management. Several open source document management systems have come onto the market in recent years.
• Business Intelligence. Another area where strong products are emerging.
• Customer Relationship Management. Open source isn’t as mature as some proprietary systems, but it’s good enough for many purposes. Where CRM overlaps social media, open source often takes the lead.
Proprietary software may still have the edge for highly demanding transactional applications, but open source can compete in a growing range of domains.
Adoption is Growing
Open source isn’t a panacea, but it certainly has attractions. Is this translating into adoption across the public sector?
There are some high profile case studies. Whitehouse.gov uses Drupal, an open source CMS. The Metropolitan Police and the CIA also use open source CMS (suggesting they’re happy about security). Several local authorities have made public commitments to open source. There’s a vocal community of open source users within schools and universities. But it’s hard to see whether such case studies constitute a trend.
This highlights another quirk of open source: being more community-based than proprietary software, it can be hard to gather comparable statistics. Traditional measures of market share, based on licence revenue, don’t apply. So we’re left with analyst surveys. These show a number of trends:
• Coverage is growing. Most analysts now cover open source. Likewise, many consultancies have open source practices. These firms wouldn’t be moving into open source if they didn’t see demand for it.
• Most analysts believe adoption is growing. This is now an almost universal opinion, at least regarding corporate adoption of open source. It also applies to the public sector in many European countries.
• The UK public sector is lagging. Another widely held opinion: the UK public sector is perceived to be lagging both other European countries and UK corporates.
What are the Barriers?
I suspect open source adoption is being held back by some systemic issues.
For a start, open source thrives in a knowledge-based economy rather than a purely financial one. A proprietary product dies quickly if it achieves few sales; an open source product can live for a long time on the enthusiasm of a small community. So market-based filters are weaker. Likewise, few open source products invest in brand-building. This places the onus on purchasers to build their own knowledge.
Open source also operates a different business model to proprietary software. It gains revenue from integration and support rather than licence fees. It unbundles elements like pre-sales support. It pushes revenue away from the centre (the vendor) out to a diffuse network of, typically small, integrators. Procurement processes that have grown around proprietary business models are ill-matched to this different model. For example, there is little incentive for open source vendors to get their products onto public sector catalogues: they get no direct revenue from product sales. Likewise, small integrators tend to avoid bureaucratic public sector procurements.
Stemming from this, procuring open source may require more work from the purchasing organisation. Without pre-sales support, purchasers need to undertake more analysis themselves. They need developers to conduct pilots. They may need to actively encourage small integrators to bid for implementations. All this needs to be managed within the bounds of procurement law. Yet procurement officers often lack the time and skills needed to do this.
So it would hardly be surprising if procurement officers are reluctant to consider open source options. Open source does indeed offer attractions to the public sector. But until these bottlenecks are addressed, it’s probably going to remain confined to the pockets of enthusiasts who are prepared to learn about it for themselves.
About the author
Graham Oakes helps people untangle complex technology, relationships, processes and governance. He can be contacted through www.grahamoakes.co.uk or at graham@...
[Non-text portions of this message have been removed]