SAP has been a expert in business applications since many years. There are more than 33.200 companies running on SAP business applications. Dr. Zencke stated that SAP has been providing more than 25 industry solutions and is counting 12 million users. Among the customers are Airbus, IBM, Microsoft, JPMorgan, Siemens, Sharp, Philips, Nokia, and many other companies.

SAP is offering tools in areas that address analytics, financials, human capital management, operations with value generation, operations support, corporate services, solution and integration platforms. SAP is also hosting controls for more than 16.000 servers and is running more than 5000 SAP and customer systems. It controls the back-up volume of more than 100 TB/day.

Dr. Zencke went on to sketch the SAP development landscapes. There are a large number of subsystems with

complex component delivery dependencies. SAP has more than 1000 concurrent developers and aims at developing one solution to consolidate into. It also has an automated test environment.

The speaker has witnessed three generations of enterprise computing which we can address as the real time computing generation, the integrated and yet distributed enterprise, and the global enterprise in value networks.

The first generation presented packaged applications and real time computing as technology innovations. The business drivers were automate corporate functions. As limiting factors Dr. Zencke cited the hardware capabilities of CPU and memory and the batch window.

The second generation trived on PCs, relational databases and LANs with a three-tier client/server architecture and ERP and satellite components. Business process reengineering was the main business driver.

In the third generation internet, portals, exchanges, and mobile devices are the novelties. The business drivers

here are business networking: everyone is a user or customer and everyone serves his customer and we are seeing competing value chains. As limiting factors, companies have to address the growing complexity and TCO.

In general, IT cannot keep up with the speed of change in business strategies and relationships, according to the speaker.

Dr. Zencke asked himself whether business computing is a stupid form of computing in comparison to supercomputing. In supercomputing the calculations are complex, the algorithms challenging, the datahandling low. Data storage is done with a main memory instead of in a database, the scalability is important and the number of users is limited. As far as memory is concerned, it is given versus critical in business computing.

The speaker showed the published results for SD benchmarks in a 3-tier internet architecture: between 1993 and 2006, the fully business-porcessed order line items per hour are going up. There is a trend back to main frame computing. Over the years we have seen the trends ranging from central to local computing and now back to central computing.

Dr. Zencke gave a few examples for throughput requirements. In mySAP SCM, the customer has 4,5 million characteristic combinations and 256 GB memory in live cache. In SAP for utilities, there are 25 million business partners and 85 million service and sales orders per year. Adaptive computing with SAP NetWeaver provides a way to virtualize application services and a single central point.

In HPBC the SAP advanced planner and optimizer is used for short- and long-term planning, addressing complex algorithms and large volumes of data. SAP liveCache Technology has a main memory base and its application logic is performed where the data is located.

The speaker gave the example of DaimlerChrysler, covering engines, transmissions and axles. There are more than 300 variants for engines, transmissions, axles, more than 400 components per variant, and more than 100.000 orders per 9 month planning period.

Another tool in HPBC is the SAP Business Intelligence Accelerator to increase data volume and shift towards real time analysis. Here we see a growth in user adoption. The SAP BI Accelerator is an in-memory technology and uses parallel processing.

Dr. Zencke gave an example for a typical BIA Installation, namely a multinational chemical company. It has more than 200 users, more than 2 terabytes of actual data, and more than 1000 aggregates. The infrastructure for this comes down to 10 x IBM blades each with 2 Xeon CPUs and 8GB RAM.

Each "next generation" introduces technology innovations and achieves that more peolple interact with the solution. Business drives application evolution. As a result we see the rise of enterprise resource planning instead of single databases, intra-enterprise co-operations instead of distributed processes in-house, collaborative business instead of collaborative processes, and virtual organisations instead of transient collaborative processes.

The fourth generation presents Enterprise SOA with web services, integration platforms and smart, intelligent devices as technology innovations. The models are becoming the business language of enterprise SOA with

business processes and their orchestration, business process components and their eventing, and business objects services.

The Mega Tenant @ SAP NetWeaver solution is making IT invisisble with SaaS. It provides virtualization of computing and storage processes and an application lifecycle management across the landscape. Furthermore, it is offering choices for flexible deployment and has the lowest switching costs between hosting and on premise.

Dr. Zencke also addressed semantic technologies. There is one thing we have to keep in mind in the non-scientific world: system architectures need to fit into business apllication context. In business, systems are mission critical for thousands of users in day-to-day work. Their interoperability is a necessity for networked business while TCO is the key business concern.

The speaker talked of technology paradise lost where he stated that we should spend less to get more approach. There is a global move toward IT simplicity and IT sandardization is a mega trend. Centraliszed computing is back in style and software as a service will surely become an option.

The new challenges to address are real time concerns, leading up to a confluence with Telco Systems, multimodal user interaction, and smart items for real word awareness. The data volume requires new query and data patterns with guaranteed response time of the services delivery within a SOA solution and accelerated analytics for the masses.

As items for a research agenda, Dr. Zencke cited the resource management where we have to take network topology into account, as well as databases and interference engines and long-lived resources. The basic operations are integration with existing work flow managment systems, automated deployment and software lifecycle management.