Adaptive Computing – Breaking Down Rigid Architectures

November 28, 2005 by admin

One area that enterprises are devoting more and more attention to is the optimization of IT landscape and system maintenance and administration, and the current discussion is focusing on key terms such as return on investment (ROI) and total cost of ownership (TCO). One of the most important factors for success is a company’s ability to adapt rapidly to changing customer, vendor, and partner requirements. This applies not only for the company’s employees, but also for the software applications it needs for day-to-day business, as well as the underlying IT infrastructure.

Modifications are becoming the rule

In recent decades, hardware and software technology has undergone significant developments. In the case of computer systems, for example, we have seen a change from monolithic mainframes to a modern client/server architecture. In terms of software, object-oriented programming languages are increasingly winning through.
These changes have had a direct effect on the technologies used in computer centers. While a few years ago, the connection of servers to storage system was based on a widely used standard (direct attached storage, DAS), there are now a large number of different connection options, for example storage area networks (SAN) or network attached storage (NAS).
However, even these versions of a system and landscape architecture are still very static. Once a customer has implemented a solution and started up productive operation, the IT landscape is usually only significantly modified in exceptional cases, or structurally changed through upgrades, enhancements, or migrations. This only allows for a limited degree of ongoing modification to meet ever-increasing demands – in relation to performance, for example.

Virtualization through SAP NetWeaver

To solve these hardware problems and enable the customer to react quickly in the case of insufficient performance, for example, SAP NetWeaver offers Adaptive Computing, a technology that makes it possible to move individual parts of a solution flexibly within an IT landscape. Adaptive Computing contains four blocks: Computing, network, storage, and control.
The Adaptive Computing Controller enables SAP customers to manage the available hardware resources dynamically and assign them to the various solutions and applications. One example: At the end of each month, a SAP customer executes various batch jobs or reports in a production system. This puts higher performance demands on the servers used. At the touch of a button, an application system can be “switched” from a 2-way to a 4-way or 8-way server in just a couple of minutes, relieving administrators of complicated and time-consuming configuration work.
To achieve this dynamic assignment from a technical perspective, software elements such as the operating system, database, and applications that want to access the resources must also be configured or programmed accordingly. A new abstraction layer is implemented in order to make the software compatible for adaptive computing. This layer removes the fixed coupling between the application and the physical server. Virtual IP addresses and logical host names ensure that the user has the same view, regardless of the hardware actually used.

Partner solutions for adaptive computing

Partner solutions for adaptive computing

SAP works closely with platform partners to ensure that the usual operating systems and databases within the SAP environment meet the future requirements of adaptive computing. To this end, the SAP Adaptive Computing Council was formed in October 2003 as a forum for ongoing information exchange. One outcome of the forum is the Adaptive Computing Compliance Test, which customers can use to find out which partners offer a solution that meet the requirements of adaptive computing. A test such as this consists of several steps, including a successful demonstration of the three functions start, stop, and relocate. A large number of further compliance tests have been carried out since the first one in May 2004. Registered users can access the current version on the SAP Service Marketplace.
In April 2005, SAP shipped the Adaptive Computing Controller to customers, and in November 2005, version ACC 1.0 SP8 of the Controller was enhanced with additional functionality:

Configuration of logical and physical system landscapes

The individual components of a landscape can be grouped into pools at both logical (application-related) and physical level. This makes it possible to split the landscape into a hierarchy, which leads to a far better representation, particularly for large landscapes.
At physical level, both individual servers and virtual units can be managed, which produces a high degree of flexibility. Mixed landscapes can be run with different hardware variants and operating systems.

Starting, stopping, and moving components

The Adaptive Computing Controller allows the administrator to assign parts of a SAP landscape, such as databases or applications, to the available servers easily and flexibly, and to move software between servers in the shortest space of time.

Recording in an adaptive computing landscape

With the help of the Controller log, it is possible to monitor which part of the application (“application service”) was assigned to which server and when. This information can then be used for internal reporting and cost settlement, for example.

Time-controlled action

The task planner in the Adaptive Computing Controller makes it possible to schedule the starting, stopping, and moving of components in a SAP landscape down to the second, enabling optimized planning for the deployment of the available resources.

Mass operations

Adaptive Computing Controller

Adaptive Computing Controller

The functions for mass operations enable the administrator to start several actions in parallel, in order to shut down a complete SAP solution on one IT landscape and start it up again on another. This simplifies and accelerates hardware updates or migrations when leasing contracts have expired, for example.

User exits

An external script can be used here to trigger customer-defined actions in a pre and post phase when an application service is both started and stopped, for example in order to import a customer development.

Adaptive Computing Controller Command Interface

The Adaptive Computing Controller Command Interface (AC-CCI) can be used to control the Adaptive Computing Controller from external tools. More information can be found in the SAP Developer Network. (sdn.sap.com – SAP Integration and Certification Center – Adaptive Computing)
The Adaptive Computing Controller is freely available to all customers who have a mySAP Business and mySAP Solution license. The software can either be obtained from the local sales channels or downloaded from the SAP Service Marketplace.

Roland Wartenberg

Roland Wartenberg

Georg Dittmar

Georg Dittmar

Tags: , ,

Leave a Reply