Re-inventing SAP

At Sapphire, the SAP users’ annual conference in New Orleans this year, Hasso Plattner, chairman and co-founder of the enterprise software giant,

 
 

Unbound: The componentisation of SAP

The ‘opening up’ of SAP’s enterprise applications suite has been seen as a gamble by many industry watchers – and certainly, it is a huge project. But it is one that has been underway at the German software giant – in one form or another – for the best part of a decade.

Hasso Plattner, the former CEO and technical visionary of the modern SAP, announced to a sceptical audience in 1995 that SAP would be opened up and broken down into smaller pieces.

In the event, SAP did make it possible for parts of the monolithic program to run separately, but the interfaces were proprietary, and the binding between the programs was tight and low level.

Later, the B2B revolution of 2001/2002 inspired SAP to open up its software still further. Its MySAP.com product (the ‘.com’ was later dropped) allowed for non-SAP functions, such as Microsoft productivity tools or B2B exchanges, to be tightly incorporated into the user experience, just as if they were actually part of MySAP. But again, a lot of this involved hard-coding links and functions into each participating system.

The 2001 takeover of Top Tier, the US/Israeli portal software company, provided SAP with a key part of its ‘patchboard’, as well as bringing Shai Agassi, widely seen as a new technical visionary, to SAP. But of equal or greater importance is the emergence of standards such as web services, Microsoft’s .NET and Java Messaging Services.

Forward to 2004, and the emerging, service-oriented business application that SAP and most analysts say is going to be so important will still consist of a complex mass of intertwined connections. The difference is that these interconnections will be more finely grained, and easier to connect or cut. It will be less of a Gordian knot and more like an electronic patchboard, with each virtual plug identically shaped into a web service.

For most SAP customers, this patchboard will be NetWeaver, now supplied as a component of MySAP. Others will use patcboards crafted elsewhere, such as IBM WebSphere and Iona’s Artix.

Before it can all work, SAP must first open up and redesign the hundreds of application functions as services. For SAP, this is (at minimum) a five-year project, which it is nearly halfway through. For customers, who have customised R/3 ERP or MySAP, and added bespoke functions and third-party programs, it will be similarly difficult.

 

 

described an experiment he had conducted.

Earlier this year, he assembled a small group of programmers and asked them to build a tricky, cross-functional invoicing application based on SAP software. But he added a condition: they were only to use standard, customer-facing tools, such as the SAP portal, business intelligence, forms integration, Microsoft Exchange and search tools. There was to be no cheating, no using shortcuts or underlying SAP routines, languages or interfaces.

Although they complained about this unnecessary focus on the user, the developers nevertheless managed to build a good, innovative application. Plattner’s point had been proven: that SAP programmers had come to realise the power of the user-facing tools now available.

Plattner wanted to demonstrate the new SAP vision of where the software world is going. It is a world where powerful new applications, involving all the functions of SAP’s underlying software suite, are built in days or weeks; where business people, not the technicians, have control of these; and where functions from many applications are routinely combined. “An application will no longer be completely inside [Microsoft] Office or inside SAP financials. There is a high likelihood that any application we build will touch many of these applications,” said Plattner.

The vision, of course, is not confined to SAP. Almost the entire software industry is now buzzing with the concepts such as the ‘service oriented architecture’ (SOA or Enterprise Services Architecture – ESA – in SAP’s parlance), composite applications and business process management. All of these, it is hoped, will give the customer power over their software and their business processes as never before.

But at SAP, this dream is now driving strategy in a way that, in public utterances at least, is not matched by any of its major competitors. “We believe in this so much that we have bet the company,” says Shai Agassi, the executive board member and former CEO of SAP acquisition TopTier, who is now behind much of the thinking.

The extent of this commitment is evident not just in a continuous string of product announcements and executive pronouncements, but in SAP’s research and development spending. It jumped nearly 10% in 2003 to $1.2 billion, and some 900 new staff have been added, bringing the headcount in R&D to nearly 9,000.

A good number of these are now working on SAP’s grand ESA project: to expose underlying functions and data from deep within SAP’s enterprise resource planning (ERP) systems, and to make them available as services which can be plugged together – along with those from other applications. (see box, ‘Unbound’). In this way, customers and partners can build new, composite applications (xApps is SAP’s term), as well as find more flexible ways of working with existing applications.

SAP has gone much further and faster in this direction than its rivals in business software, Oracle, PeopleSoft and Siebel.

“Oracle is moving in the same direction,” says Massimo Pezzini, lead integration analyst at research company Gartner. “And there is something in the making at PeopleSoft.” But, he adds, neither has set out with the commitment or clarity of SAP.

“We believe SAP’s direction is not only different, but it is better,” agrees Erin Kinikin of Forrester Research.

High stakes

But is it, as Agassi and many analysts suggest, a gamble? And why is SAP pressing ahead so much faster, and more radically, than its rivals? One thing is clear:

 

NetWeaver: Not for Dummies

Even if they chose not to take their complimentary yellow paperback home with them, visitors to this year’s Sapphire SAP user conference could not have missed the marketing message: By commissioning tens of thousands of copies of NetWeaver for Dummies, SAP is telling the world that it is pretty serious about this product.

NetWeaver lies at the core of SAP’s big Enterprise Services Architecture (ESA) migration. “It is the foundation of everything we do at SAP,” says Shai Agassi, the former Top Tier CEO and SAP board member who now heads the NetWeaver business. That is because the new service-oriented version of MySAP will not function without it (see box, ‘Unbound’).

The launch of NetWeaver in 2003 sent ripples through the software infrastructure business. The product currently consists of a portal builder, an application server, a master data manager, some business intelligence and analytics tools, an integration suite and various other tools.

Because of its comprehensiveness, NetWeaver appears to compete with all manner of application servers, integration and portal products. In other words, it appears to signal SAP’s intention to move out from its application territory and become an infrastructure supplier.

Most existing users won’t pay extra. NetWeaver is pre-built into the latest versions of MySAP Business Suite. SAP thinks users with NetWeaver are more likely to buy and integrate new SAP components.

Despite some analyst criticism, the consensus is that NetWeaver will be a vital component for almost all users of SAP products – and ultimately, a significant product for many others too.

 

 
 

SAP, in spite of its apparent willingness to consider a takeover offer from Microsoft early in 2004, is not doing this from a position of weakness. By almost every metric, SAP’s domination of the high-end software market is growing.

In mid-2004, for example, SAP’s market share in ERP, measured against a narrow line-up of large competitors, grew to a record 55%, up from 51% a year earlier. Its recent financial results confirm the trend: for the first half of 2004 to June, sales rose 6% to EU3.3 billion, with net profits up 18% to E478 million.

If this is really a gamble, then this is some stake to put up. But there is another way to look at SAP’s move: that the commercial risk for the company is greater if it stands still – or even moves too slowly. That, anyway, is what executives at SAP have concluded.

The reasons are many. First, in spite of some localised revenue increases, the overall trend in recent years has been for big organisations to spend less on big software packages. Overall new licence sales growth for high-end business applications software has been slow, flat or falling for several years.

SAP, with its pole position, has fared much better than others, but its recent licence growth spurt is partly due to a wave of new products coinciding with customers’ IT budgets being freed up, and the relative weakness of some competitors. Overall, application licence growth remains subdued, and the danger is that this will translate into slower services revenues.

Customers have spelled out the reason why sales have slowed in no uncertain terms: ERP systems are still perceived as expensive, and difficult to integrate, modify and extend. New expenditure is difficult to justify, they say.

To demonstrate the point, the vast majority of all ERP customers are using versions of the software that are three or more years old. In SAP’s case, some 60% of all customers are still running R/3, which was replaced by MySAP back in 1998.

Pre-emptive strike

Another looming problem: For years now, Microsoft and IBM have been telling customers it will soon be possible to build many of the integrated, cross-functional applications they need without investing in major new packages. By plugging together existing and new services and components, and using some technology and services from these companies, they will be able to

 
 

Q&A: Henning Kagermann, CEO, SAP

Information Age (IA): How long will it take to turn MySAP into an open, service-oriented application?

Henning Kagerman (HK): It will take us two to three years, not five to ten years [as some have said].

IA: Gartner speaks of the end of the monolithic application. Do you perceive this is as a threat?

HK: For us, it’s not a threat. We have no monolithic software. We started very early to make the shift: we have a lot of components, with the software running on top. R/3 [the enterprise resource planning suite that preceded MySAP] was viewed as monolithic – even that was not true, but it was certainly more monolithic than MySAP.

In the 1990s, there was big pressure on functionality: functionality was low and performance was bad. If the market is driven by functionality, tight integration is the answer. But now, companies are looking for flexibility. [They want] a modular, component-based architecture.

IA: How important are xApps to your business? Are they a way for customers to make better use of SAP, or do you see them as a major business opportunity in their own right?

HK: Component applications are being built using other components and services. Some of those are xApps; others are replacing existing functionality with flexible, service-based functions. From an architectural point of view, xApps are composites.

IA: How would you compare NetWeaver to, for example, IBM’s WebSphere or BEA Systems’ Weblogic? Is it aimed at non-SAP customers?

HK: In our case, using the network to get to the clients is not the main purpose. Our drive is to help clients put it all together, with all products running on NetWeaver. The NetWeaver platform is extensible, and you can use it to connect non-SAP systems or develop new add-ons. It is not pure middleware.

IA: How do you assess your partner Microsoft as a competitive threat?

HK: [Microsoft] is tough competition at the low end; it is not a threat at the high end. I am not aware that Microsoft wants to enter the high-end with applications. .NET is also a platform, but it is not a threat at the high end.

 

 

get much of the flexibility they need. As Plattner’s experiment showed, they will soon be proved right.

All of this encouraged SAP to make a pre-emptive strike, beginning work on a radical makeover as early as 2002. The strategy has been to turn MySAP into a service-oriented platform, driven by the integration hub NetWeaver, within five years.

It was not a moment too soon. By mid-2003, Gartner was warning its clients of the coming end of the monolithic application. And shortly after, its analysts began advising clients that if their supplier did not have a service-oriented roadmap, they should “evaluate alternatives”.

For SAP, of course, it is vital that customers adopt the SOA/ESA philosophy, and do so by adopting NetWeaver. This, it hopes, will encourage them to start buying new functions, even if they only migrate gradually from older versions of the ERP software.

Time to act?

By focusing on the many benefits in terms of business agility, SAP is now attempting to instill a sense of urgency among its usually cautious customers.

“You can do your own roadmap. But you have no time. You should start now,” Kagermann told customers recently.

Few people are under any illusion about the ease with which this vision will be achieved. This is, after all, SAP software, where typical roll-outs traditionally take years. “You have to go through the process,” says Agassi. “It’s work that needs be done very diligently. We’re asking you to really think big.”

What do customers and partners make of all this? In spite of calls to suspend their disbelief, many are still wary, not least because they know – and have been warned – that it will not be easy. “I do think it’s important, but it will take a while before we know which way it will all go,” says Eric Sol, business technology manager at Rohm and Haas, the large speciality chemicals company and a NetWeaver and MySAP user.

“The whole concept has been favourably received. But most SAP customers are still struggling with more basic problems with their ERP landscape,” says Paul Bray, head of the SAP practice at IT services company Atos Origin.

This point demonstrates clearly enough SAP’s problem. It has invested heavily to get ahead and tried to solve some pressing business problems, using its advantage to steal some market share at the same time. If it can do this, it can avoid any slowdown in new sales.

But customers, and developers, are not entirely convinced that the time is right, nor that the roadmap is clear. In fact, says Pezzini of Gartner, SAP’s first big challenge will be to convince its own developers to commit to the vision.

SAP, however, is not lacking in evangelistic zeal. “This is a great moment for the industry,” says Agassi, adding that customers need to suspend their disbelief in order to get the dramatic benefits. You can start small, he says, but “think big”.

Avatar photo

Ben Rossi

Ben was Vitesse Media's editorial director, leading content creation and editorial strategy across all Vitesse products, including its market-leading B2B and consumer magazines, websites, research and...

Related Topics