The Hindu Business Line : `Accent' on trends

Accenture, the consulting-cum-software heavyweight, shares insight on key trends that are set to dominate the technology landscape in the next five years. eWorld zooms in on the road ahead.

The key question at the R&D labs: "What new things are achievable that were not achievable five years ago, in the light of these trends?"

Recently in Sophia Antipolis, France

In a globalised environment, one of the toughest challenges facing the Chief Technology Officer of any software giant is to identify technology trends that lie at the intersection of business and technology.

And if it happens to be the CTO of a $16.6-billion consulting-cum-software outsourcing heavyweight, the complexity multiplies several times over.

Meet Donald J Rippert, Chief Technology Officer of Accenture, who grapples with complex issues of this kind on a day-to-day basis.

In an exclusive presentation to a group of visiting journalists from India at their Technology Labs in Sophia Antipolis, said to be something of a hub of `Telecom Valley' on the French Riviera, Rippert outlined some key trends that are likely to reshape the technology landscape over the next five years.

Confluence of hardware

In Accenture's vision process, which is the beginning of their innovation cycle, Rippert says, they look very hard at fundamental technology changes.

At this point in time, Accenture is placing big bets on the confluence of hardware trends that surround storage, display, bandwidth and processing power. "All these technology trends are growing almost vertical as far as improvements in price-performance are concerned. There comes a point in time in this curve, because it is cumulative and logarithmic, when the improvement in all these years put together is significant. We are getting to a point where, I believe, the fundamental technologies are outpacing the industry curve," he says.

In the 2000-01 timeframe, in the adoption of browsers and PCs, industry had caught up with technology. Technology was being utilised nearly to its capacity. The same trend is slowly beginning to happen in the fundamental technology space.

As the gap between the applications set that is available in the marketplace and the capability of technology starts narrowing, it can create opportunities for a lot of different companies. The companies are slowly pushing the limits in the case of chips, storage, etc.

The implications of these grinding trends are that in storage, the price of storage per gigabyte has dropped from $50 in 2000 to $1 in 2005. The Accenture prediction is that storage costs will drop to $0.01 by 2010. And this will set off a truly revolutionary trend in storage — from being prohibitive to making costs totally irrelevant.

In the case of sensors, the total worldwide market for sensors was about $100 million and the average selling price was $25 each in 2000. Five years later, the market size has grown five times, with the price coming down by half per unit to $12. If we shift the clock to the next five years, the market is expected to grow to $1.1 billion and costs are likely to drop to $5 each.

The trend in wireless bandwidth is somewhat similar, with 3G slowly becoming a dominant technology. And in display panels, just like Moore's Law for chips (which said that the power of the processors will double every 24 months; it turned out to be 18 months finally), Kitahara's law for display panel states that every three years, as the screen area goes up by 44 per cent, the weight and panel thickness will decrease by a third. The multi-tasking possibilities with multiple screens will open up as prices of display panels drop (See Table).

Imagine how the application sets can change when the sensor costs drop dramatically, say in a camera phone, wireless bandwidth increases and there is a virtual collapse in the cost of storage, which essentially means that one can store as many pictures as one wants.According to Rippert, it repeatedly brings them to the key question at the R&D labs: "What new things are achievable that were not achievable five years ago, in the light of these trends?"

As an example, he cites a predictive maintenance project that Accenture is doing for a US-based utility company. Using the advancement in sensor technology, Accenture has been able to monitor turbines based on specific attributes and highlight failure rates that can alert utilities to proper shutdowns.

The search for `enterprise search'

Even as the opportunity landscape widens, it presents a set of new challenges for companies such as Accenture. There is a growing apprehension that companies could be buried in an avalanche of data.

The implications are that chipsets will become more powerful. By 2010, there will be eight times more non-PC computers than PC. But this will pose a key challenge of integrating non-PC based processing devices, which is far more difficult than PCs.

Estimates also show that the number of Web pages has grown from 200 million to 12 billion, an increase of 60 times in less than 10 years. So, filters to manage content will become more important than ever before.

It is also projected that the average sizes of customer data warehouse in data-intensive industries have grown six-fold in the last five years. A lot of databases do not operate effectively with that kind of scale.

"Corporations have very poor search capabilities to navigate through this mass of data. They get a much better search on Google than they do on enterprise search," adds Rippert.

Clearly, this means that `enterprise search' will be the next holy grail of technology.

Anything that Google or the less widely known Scandinavia-based Fast, a developer of enterprise search technology (recently adjudged by Forrester as a leader in Enterprise Search platforms) can do to bridge the gap between application set (effective database search) and technology (search engine capability) can revolutionise this area.

Bogged in `ocean of code'

The number of developers and the language in which they develop code has swelled over the years. Strangely, despite advances in technology, Cobol mainframes still play a significant role in application development. To highlight this point, Rippert says that Cobol mainframes still "account for 75 per cent of all computing transactions, 90 per cent of all financial transactions and 150 billion lines of code worldwide is still growing between 3-5 per cent a year."

"The number of professional Cobol developers is going down, while the number of lines of Cobol code is going up. Fewer programmers generate a lot more code and the problem is that there is just too much code. Assume there is an application that does Function X and you want it to do Function X.2. Today, you can basically regenerate another application, have it do it and call it Application Y. The problem is that a company runs Application X and also the new application. This means that you are developing more code than the functionality that the system is giving. The implication of this on technology — tools to the developers with this amount of invented code are developing so fast and so large that they are beginning to stop the ability to add functions. Companies are getting bogged down down in an ocean of code."

The big area of opportunity for companies such as Accenture is to start retiring code. Not just regenerating but retiring code. "If a corporation has a million lines of code and they plan to add 10 per cent additional functionality next year, we would like to put 2,00,000 lines out of commission and add 1,00,000 lines. Thus slowly we have a lower code base to add to and lesser redundancy, going forward."

Quality of IT spends

A survey conducted by Accenture of 400 global organisations (each with an IT Budget running into $0.75 billion to $1 billion) shows that high performers have significantly lower programming fixes compared to low performers. Instead, the high performers spend more on enhancing, integrating and building applications and adding more functionality to their application set.

"The more efficient the company is, the more it will spend on adding new functionality while the less proficient one will do that on fixing existing code." This survey points out that high performers spend on an average only 5 per cent on fixing existing applications compared to 16 per cent spent by low performers.

This straightaway opens up a significantly higher budget for enhancing, integrating and building applications. Typically, high performers spend 45 per cent in this area, almost 12 percentage points higher than the low performers.

Companies are also increasingly using virtualisation or partitioning through VMware or Egenera, which let companies save on server and hardware costs. Offshoring data centre operations is also saving a lot of money for companies, which is going a long way towards improving the quality of IT spends.

SOA and Web 2.0

Of the 150 technologies analysed by Accenture, services-oriented architecture or SOA (the technology that underpins Web services) ranked as the top technology with the greatest impact on corporations in 2006 and it is expected to stay in that slot in 2007.

Showcasing the power of SOA, Rippert says that in the area of healthcare, Accenture is attempting to make e-prescriptions work. It is an extremely complex process as hospitals have one platform, doctors' offices and the pharmacy work on another, the healthcare provider has yet another and the insurance provider, one more. These are five different entities for which Accenture is planning to provide a link through SOA by generating composite applications that can mesh together.

Web 2.0, the new generation of Web applications that are more dynamic and participative, is slowly changing the face of the Web. Mini Cooper, a sports small car manufacturer, originally of British origin and now part of the BMW group, has used Web 2.0 to direct 50 per cent of its leads through its Web site and almost 30 per cent of its sales now take place through the Web. JP Morgan has also used Web 2.0 in the derivatives arena and Salesforce.com has also clocked vast improvements using this technology.

Watch this space for more applications from the R&D Technology Labs of Accenture.

Fundamental vs aspirational

Sometimes making technology bets boils down to what Rippert calls a distinction between `fundamental' and `aspirational' change.

"If you look at the core technologies, we are analysing what is happening at the fundamental level in processing, in storage and display and making some bets on fundamental technological change."

We believe that "if you bet on fundamental change, you make more good bets than bad debts. But if you bet on aspirational change, you make more bad bets than good."

For instance, if you go back and find out why `client server' worked so well, you will find that it was because it took advantage of several fundamental changes in the client's processing requirements relative to performance and great improvement in bandwidth. Once those trends hit a tipping point, it took over the mainframe requirements of companies across the globe.

Contrast this with artificial intelligence. Even though it showed great promise, why did it not live up to expectations the way client server did? Because it was not a fundamental technological change, it was largely aspirational. "People said it would be great if we can write software that can operate the way people do. We can encapsulate the work of the knowledge engineers and once they retire we will not lose all that knowledge. Great goal, but no fundamental technology change to assert this logic", adds Rippert.

Business Implications

Donald Rippert

Following the presentation, eWorld posed a few questions to Donald J Rippert on SOA and its business implications.

Do you think that services-oriented architecture (SOA) is a compelling value proposition for the technology market at this point in time?

What worries me is that it (SOA) is more aspirational rather than fundamental. Major technology vendors can say: Why should I give up my differentiation? But others are also saying: It is so expensive to implement ERP, but SOA makes it a lot less expensive. Everyone is better off. SOA is more of a co-operation around standards than anything else.

Has SOA become an industry standard, because SAP and Oracle have realised that the ERP spending cycle has come to an end and they needed something new to build up demand?

The original drivers of this were Microsoft and IBM. Microsoft saw too many companies saying: I do not need an upgrade. This was as long as there was middleware for an application that worked. Each of the different groups finally came to the conclusion, very sincere all of them, whether it was Microsoft, Oracle or IBM that there was a question behind 75 per cent of all IT budgets. Out of $100 of budget, $65 was new functionality. The whole development of systems software has been driving the vendors to go towards a standard. That is how the economic impetus and theory on SOA has evolved.

The banking industry is having a very hard time of getting the operations of Bank X to work with Bank Y. Say Bank X is running Oracle, while Bank Y is running mySAP. So, banks are saying, `you get the synergy savings and I will get the businesses to drive the upgrade.'

Capital market guys are trying to make it a faster way to implement products, say for trading or derivatives.

maverick@thehindu.co.in