Big players take steps to realise the dream of systems that diagnose and fix themselves

Self-healing systems promise to take some of the manual labour out of IT.

Self-healing systems promise to take some of the manual labour out of IT.

IBM, Microsoft and CA have all hit milestones on the road to enabling self-healing IT systems.

Automated monitoring and self-healing systems promise to lower costs, reduce complexity, and take some of the manual labour out of managing heterogeneous ITsystems.

The technology is designed to predict, diagnose and fix problems before or as they occur, and to dynamically allocate computing resources as and when they are needed.

When applied successfully this technology could mean that IT failures have little impact on business processes, and that customer data will not be lost when a system freezes. It could also mean that by using dynamic allocation of computing resources, spikes in demand will not slow down the system.

Updating software automatically is not new. Operating system suppliers such as Microsoft and Apple use it in their software, as do suppliers of anti-virus and other utilities. But system self-management technology takes this a step further, linking whole applications and hardware subsystems.

Microsoft started its research in the early 1990s into what is now called the Dynamic Systems Initiative (DSI). At the same time, IBM was working on its eLiza project, now known as the IBM Autonomic Computing Initiative.

Despite the advances in self-healing from IBM, Microsoft, CA, and moves by Sun Microsystems and numerous specialist suppliers, analysts have warned that the technology is still in its infancy, and urged IT managers to be cautious about purchases.

Carl Greiner, senior vice-president of infrastructure at Ovum, said although system self-healing is at the early stages, it will be an essential technology as systems became more complex.

"The big push now is to have IT systems that automatically provision and correct themselves. For that we have to go in and look at our environment, monitor it and determine where the fix has to take place. This is not easy stuff. We can view the environment superficially, but can we fix everything yet? No. But I think we're heading in the right direction."

Frank Gillette, principal analyst at Forrester Research, said, "The industry is not very far down the track on self-healing technologies.

"Think of how far cars have come in 100 years and then compare that to how fussy and unhelpful computers are when they have problems.

CA's self-healing aspirations

CA plans to add self-healing capabilities to its software that will automate Unicenter administrative tasks.

In January, CA bought Control-F1 which develops support software that can prevent, detect and repair end-user computer problems before they disrupt the wider IT system.

The new CA self-healing tools will automate the detection, diagnosis and repair procedures for remote users. It will also use software agents to monitor registry and configuration settings and the state of applications and hardware, automatically applying fixes as necessary. Employees will also be able to resolve computer problems via the internet or intranet.

CA has already integrated Control-F1's products into its own range, and plans to sell them in standalone form as well.

IBM follows autonomic computing path in Tivoli series

IBM has introduced three autonomic management products in its Tivoli series. The software marks the next wave of self-managing, autonomic technology from IBM by proactively healing technology problems.

Nick Drabble, IBM's automation business manager UK for Tivoli software, said, "In essence we are trying to understand where outages will occur and prevent them."

According to Drabble, at the core of a self-healing system is the ability to automatically find out what the IT system contains. There is also a database that can store information about all aspects of the IT system and its dependencies.

IBM's self-healing technology includes IBM Tivoli Monitoring version 6.1 (ITM). The systems management tool can report on the health of an IT infrastructure, including its applications, databases and hardware.

ITM integrates tightly with mainframe tools and gives a single view of an IT infrastructure, reporting in real-time what is happening in different parts of the system, said Drabble.

The self-healing aspect of ITM is based on a correlation engine. "This can analyse potentially hundreds of error messages. It can pinpoint the root cause, and also suggest the right action to take. It gives expert advice and guidance, and you even have the potential to have that advice executed automatically with a predefined workflow," he said.

ITM also uses self-healing to keep web-based applications running. It can automatically ease bottlenecks caused by e-mail or bill paying systems by allocating network and server resources or server memory as they are required, and according to predetermined procedures. IBM has tested the software over seven months in nearly 100 organisations.

This year IBM plans to release a change and configuration management database which will track the status and activity of all elements in an IT system, their real-time activities and their interdependencies. The aim will be to manage and automate changes to the IT system.

Another IBM product currently being used to provision applications is Intelligent Orchestrator, which can automatically provision datacentre resources according to predefined workflows. "It can take many man hours down to minutes," said Drabble.

Microsoft's dynamic move

Microsoft has based its self-healing and self-managing systems initiative on the ability to track the health of heterogeneous IT systems.

Self-healing is part of its Dynamic Systems Initiative, which will see a range of products incorporating some degree of self-healing technology, including Windows Server Update Services, Microsoft Operations Manager (MOM) 2005, Systems Management Server 2003, Visual Studio 2005 and Virtual Server 2005.

Systems Center Capacity Planner 2006 is Microsoft's first foray into system self-healing, said Alfred Biehler, Microsoft product manager for management products in the UK.

The system was released in December 2005 and is designed to help IT managers automatically plan and manage the hardware and software capacity of their Exchange 2003 and MOM 2005 servers.

It also reports on and manages performance trends and bottlenecks, in a similar way to IBM's Tivoli ITM.

But, Biehler admitted that true self-healing is not quite here yet, and that there was still much work to do before self-healing systems truly arrive. "In the future I would like to have a machine that never breaks and that I would not have to pay attention to - at the moment it needs a human to make decisions and to fix it," he said.

"Key to the Dynamic Systems Initiative is that we capture the knowledge of the servers and hardware configuration in a language that the capacity planner can understand.

"Nothing is available now that has all that knowledge captured. MOM is the closest and captures that knowledge of Exchange. But we have to agree on a standard first."

Forrester Research said that Microsoft had good aspirations for the Dynamic Systems Initiative, but that it will take years to fully implement.

Biehler agreed, but said Microsoft technology would soon be able to fix predictable problems and automate mundane tasks.

Read more on Business applications