One of the perceived benefits of SOA is business agility and system adaptability. In this post, I am going to discuss some aspects of this.
There is a lot of vague and woolly talk about business agility - especially from the software world, where a wide variety of software products, platforms and paradigms are optimistically supposed to have some magical effect on flexibility. But try selling flexibility to the CFO of a large company. ("Excuse me sir, would you like to buy some flexibility? It will only cost you ten million dollars." "Exactly how much flexibility do I get for ten million dollars?" "Ooo, loads and loads, honest. Look, here's a graph we've just made up. And a 2x2 matrix.")
Among other things, business agility means keeping your options open.
- Options are worth more when conditions are uncertain.
- In financial circles, the value of options is calculated using the Brook-Scholes formula. Stock options increase in value with the volatility of the underlying stock. This encourages risk-taking by executives.
- Real-Option theory applies the same financial logic to management options. See book Real Options by Amram and Kulatilaka.
In the "plug-and-play" service economy, the business may gain option-value in several ways:
- the ability to replace specific partners
- the ability to flex the boundary - moving specific services inwards or outwards across the organization boundary
- the ability to access heterogeneous domains
- the ability to change the configuration (geometry)
To go beyond bland optimism, we need to think rigorously about the business value of openness and extensibility, and how this value can be achieved by suitable configurations of organizational and technical systems - including SOA.
Technical artefacts (such as computers) often have empty expansion slots.
Ben Hyde discusses the option value of empty slots - who benefits from this unused real-estate?
When the vendor creates a slot he’s relinquishing control over some amount of value-creating energy implicit in his offering.
If there is anything consistent about how they get filled in - for example they all start sporting graphics cards - that’s not stable. The industry will absorb any consistent slot usage into the core.
The space of empty slots [... appears to be ...] a long tail [... which ...] actually has negative value. Since the majority of buyers don’t use the slots they are getting no value from them. The hardware makers are including them not because of buyer demand but because the dominant players in the market want them. The dominant players in the market want them because they tap into the value created by the generative process that all that empty real estate creates.
In other words, the expansion slots provide some temporary free space for innovation. But the big players monitor how this free space is used, and will seek to recapture any rich pickings.
Martin Geddes discusses a number of other technologies (past and present) that have added value by providing options. Many of these were initially dismissed as inefficient. Martin's list includes relational databases and XML - we might add web services.
Compare this with the adaptability of buildings.
Lots of old houses were built with outside toilets and no bathroom. Most of those that survive today have had inside toilets and bathrooms added. Many people convert loftspace to make extra rooms, or add an extension into the back garden. Old houses are often relatively easy to alter or extend, subject to planning regulations.
Construction companies build new houses on tight estates with shallow roofs - so there is insufficient space for expansion. Bathrooms are attached to so-called master bedrooms, thus constraining alternative ways of using the space. It is as if the construction companies deliberately set out to build houses that are pre-adapted to a specific domestic norm, thus forcing people to move house every time their domestic requirements changed.
See
Stewart Brand: How Buildings Learn.
The human being is a highly inefficient machine, with few natural advantages over other species and practically no innate skills. It takes many years before it can find its own food, or run half as fast as an animal. It has a large brain, consuming vast amounts of energy and other nutrients. The exceptionally long period of dependence, together with its feeble skills in relation to other animals, represents a considerable biological burden.
How on earth does this gross inefficiency survive? In complex and dynamic environments, there is a compensating evolutionary advantage - the human has lots of expansion slots - can learn new skills, including collective skills. It can develop new languages - both natural languages and problem-specific vocabularies. (DSL anybody?)
So what are the expansion slots of a business? Obviously new products, new suppliers, new business partners, new channels. But more importantly, new responses to changing demand. With a stratified model of the demand-side, and with appropriate supply-side platforms, we should be able to (re)design a business to deliver requisite levels of agility. And this will naturally include space for innovation.
Update: Stewart Brand has now introduced the term
pace layering for the principle that stratification should be based on the differential rate of change. The term
shearing layers
refers to what happens when this principle is not followed.
See also Andrew K Johnston,
Business Flexibility - An Analogy (2005)
