JP Rangaswami, CIO of BT Global Services since 2006, is a highly respected voice in the UK IT-industry, a fellow of both the British Computer Society and the Royal Society of Arts. Before joining BT Global Services he was the global CIO of the investment bank Dresdner Kleinwort Wasserstein, where he introduced a MySQL-based system. He was named CIO of the year by Waters Magazine in 2003 and CIO Innovator of the year by the European Technology Forum in 2004.
In your blog “confused of calcutta” you confess to be passionate about open source. What from the business perspective of a CIO would you say is the major impact of open source on the IT industry?
While it might not yet be evident to everybody, open source software today is mainstream. Open source software (OSS) is about commoditization. Upon reaching maturity open source components together form an infrastructure upon which to build applications and services. Since commoditization results in less differentiation it will be necessary to differentiate on the level above the infrastructure.
Just how far do you think this commoditization can reach?
Business is switching from making money with the product you sell to using it as a door opener for selling services. Talent doesn’t scale or let itself be commoditized so this is what you will pay for in the future. Human capacity should be used to create value for the customer rather than in technical details. Any service company must transform from being a product shop into looking at the functionality and fit that the customer wants – and I’m talking about the individual customer. It is particularly important for a company like BT Global Services to develop tools for reaching different types of customers with their individual needs. Strategy must be based on what the customer wants rather than on what suppliers want to sell. That kind of complexity on the other hand requires standardization in components to deliver better services. OSS is about commoditizing and it fits well into this strategy. Value lies in providing a standardized infrastructure to let us focus on creating value on top of it.
There are thousands of open source projects out there in all stages of development. How would a CIO know which ones to build on?
It is obviously important for BT Global Services to identify OSS components as they reach the appropriate level of maturity and commoditization suitable for enterprise business. There is a point in development for any provision of OSS where we can enter and then scale up in our business. The most important factors are size of the surrounding community and the maturity of the project. Eric Raymond once stated “Linus’s Law”: “Given enough eyeballs any bug is shallow”. This means that in order to achieve the superior code quality which characterizes mature open source projects, a sufficient number of independent developers have to be engaged in the project. Concerning maturity the specific project has to be a tested and acknowledged part of the OSS infrastructure, much like MySQL is a part of the commonly used LAMP-stack.
Traditionally open source entered enterprises through technical subsystems and then has grown into the core business. You however claim that OSS might be considered as a strategic choice on its own merits?
Once small systems might have started on OSS only then to be shifted to proprietary software as they grew. This is however a dated way of looking at it. I perceive two evolution stages for OSS components: They started out at an individual and personal level but have evolved to an enterprise level. MySQL for instance has clearly taken the step from individual to enterprise resource. Databases have had open standards for a while. MySQL has moved away from being mostly suited for small individual systems during later years. Historically software started at enterprise level and after a while became useful for individuals but with OSS it’s the other way around. I actually foresee a third evolution stage for databases: scaling to the entire industry segment.
So do you believe that open source at some point will replace proprietary software altogether?
It’s really not an important or even relevant issue. In time I believe open and closed source worlds will converge but living in current hybrid environments it’s even more important with standardization. Otherwise all you do is pass on the complexity to the customers. This is however not an idealistic war but a pragmatic necessity. In adopting open source software there’s of course the usual curve of enthusiastic early adopters, skeptical middle pragmatists and stubborn laggards often making a philosophical stand. Acceptance of OSS, however, is inevitable while interest is shifting from which systems are used to what value the systems create – from supplier to value adding. Our future business is to teach the customer to fish rather than sell them fishes.
Despite the advantages you describe we see that acceptance of open source software is slower in some regions and verticals than in others. What do you think is the reason for this?
I believe that incentives to use OSS might be less in markets dominated by intermediates choosing solutions for the end customers. Their revenues might to a certain extent be dependant on lock-in and infrastructure costs. To change the situation demand for less complexity and cost reduction, it must come from the end customers.
Do you see any less obvious advantages with open source apart from the commoditization?
Yes, I’d like to highlight three unsung values of OSS: 1. Change in training cost for talent resulting in a lower cost of entry for new companies. This opportunity gave birth to the whole Indian offshore IT industry. 2. Security value. Inspectable open source code actually reduces rather than increases security costs. This may seem like a paradox but is already understood by security experts. 3. Lower cost of merger and acquisition. Simple standards and non-lock-in reduced costs to merge cultures in companies or between companies.
If you are a CIO in an existing proprietary environment, should you be considering OSS for a new database purchase?
You should go to jail if you don’t. Seriously OSS is mature. You can’t find a fortune 500 company today that doesn’t use open source somewhere. The key is: the decision has shifted from talking about labels to talking about value. OSS is mainstream and MySQL can hardly be called a startup anymore.
So what’s your advise to a CIO that is curious but still hesitant about using OSS?
Make your life simpler by using OSS. The chessboard is getting crowded so where appropriate you should strive for standardization. Mature OSS can give this standardization with higher quality due to “the million eyes” and at an affordable price. Where’s the argument against it?