Dieses Bild zeigt

New at the University: Prof. Steffen Becker

“Our Laboratory is Reality”

Steffen Becker took over the post of Professor for Reliable Software Systems (ISTE) on March 1, 2017. “We analyze whether the software structures are suitable for guaranteeing the quality of the software”, says Becker of his work and the work of his team.

The amount of data is increasing exponentially and because systems are becoming more and more accessible in an everyday, non-specialized context, the current systems must be capable of meeting very different requirements. And that is why Becker has been active in this field for many years.

The Primary Concern: Analyzing Software Quality

The youthful man from Middle Hesse begins by clearing up a misconception: “We do not develop software. We are involved with procedures; in my case with the procedures involved in analyzing software quality.” Ideally, research and industrial service providers work hand in hand. “Our analysis looks at what the designers want to achieve with the software. We then analyze whether the software is really capable of meeting the requirements”.

Preventing Chaos

The title of the Stuttgart professorship does not describe the full scope of the research carried out. The Information Systems expert explains, that his work not only involves reliability, but also software quality in a range of different applications, particularly in terms of software performance and how it responds to large data loads.“ The reliability, i.e. ensuring that the software does not fail, is just one of many aspects of his research and lectures. The same applies for the topic of serviceability, meaning whether the software can be developed further or whether errors can be easily corrected. Performance, reliability, serviceability, availability, scalability, elasticity, efficiency or safety – these are the fundamental software qualities that the department is involved in, either to a lesser or greater extent.

For the industrial sector, modelling, analysis, and the optimization of software architecture are extremely important aspects. The valuable preparatory work, which ideally results in a flawless software system, can prevent chaos and the high additional costs caused by the need to upgrade and improve the software. “As software engineers, we are attempting to write a handbook that systematically depicts the requirements for using various designs. We analyze whether the software structures are suitable for guaranteeing the quality of the software.”

Developing Methods for Users

Researching methods for software developers – this is one of the department’s primary concerns. A specific recent example would be the healthcare system in the USA. “The entire American population, so approximately 200 million people, must have access to this.” Consequently, an industrial service provider was commissioned to develop a suitable system. “The developers could say: ‘That is a complicated system. Let us see what methods Becker has developed and use those.’ The result would have been: ‘if you develop it as planned, the entire system will break down.’ But because they didn’t used this method, the entire system broke down”, says the software engineer of the debacle in America.

Preliminary Analysis Instead of a Complete Meltdown

Becker has worked together successfully with industrial partners such as ABB, Phoenix Contact or IBM. “The systems had already been produced and we analyzed them.” They are mostly individual systems that the industrial partners build for their customers. For example, ABB develops software that can control the facilities used by the company BASF. Generally speaking, everything works fine, but what happens if something goes wrong? “If the pumps jam at BASF the warning light flashes and the chemicals stop flowing. This changes the mix of chemicals in the tank and the tank’s warning light also starts to flash. It does not take long until all the lights on the control panels begin to flash.”

An alarm does not sound by itself for long; it is usually rapidly followed by thousands of other alarms. This is where Becker’s analysis comes into play: it is all about making the system as strong as possible in case of an ‘alarm burst’, so that the necessary user inputs can be carried out that are required to rectify the critical situation. This ensures that that even in the case of an alarm, there are still enough reserves to operate the system. However, in America the ‘alarm burst’ prevented the system from reacting at all because of all the flashing lights.

Specific Research

“Our Laboratory is the Reality”, says Becker of his research. He teaches his students about the methods that can be used to ensure software quality. “So that they will be able to begin working directly for our industrial partners in the future.” Software architects like those in Becker’s department have been around for the last twenty years, but this way of working is not yet established in practice and in the industrial sector.

The Backlog Experienced by Industrial Service Providers

“There are an increasing number of systems that fail for reasons of quality”, states the professor, emphasizing the relevance of the work carried out by his institute. As an example, he refers to the Internet of Things (IoT): “Billions of data can enter a system each hour, because every device that is connected to the network sends data. It is no longer possible to build a system simply by testing its functions.”

While his colleague Küsters, from the Institute of Information Security, uses an overload scenario to bring a system to its knees, Becker attempts to prevent the system from ever being brought to its knees. Becker: “Of course, this is not something that scientists do alone, we work together with the developers. Only they know the system, but we know the methods to use to analyze the system.”

There are still very few experts in this field and Becker and his team’s software quality analysis has made them esteemed partners in the industrial sector. And the software whisperer, in turn, needs suitable industrial partners for all aspects of his work “who actually have the problems that we want to solve, and these can be found in abundance in Stuttgart.” The chances are high that Becker’s academic nomadic existence may soon come to an end – and not least because the native Hessian is clearly at home here.

 

Dieses Bild zeigt Becker
Prof. Dr.-Ing.

Steffen Becker

Head of the Reliable Software Systems Group

[Photo: U. Regenscheit / University of Stuttgart]