Heal Thyself

Technology that heals itself has long been the domain of movie theatres and science fiction novels. However, as a number of vendors develop their own self-healing initiatives, it may not be such a pipe dream after all.

  • E-Mail
By  Matthew Southwell Published  September 28, 2003

I|~||~||~|Although the idea of highly intelligent computers replicating themselves and taking over the world is still the domain of the Wachoski brothers and other science fiction filmmakers, a number of IT vendors are in fact developing technologies that are capable of healing themselves.

Rather than relying on human intervention, these super smart systems have the ability to diagnose, heal, re-configure, optimise and protect themselves, simply stopping along the way to inform IT managers and administrators of the course of action they selected.
On the network side, Cisco’s initiative to create a comprehensive systems-wide approach to resiliency and increased IT network availability falls under its Globally Resilient IP (GRIP) project.

For software, Microsoft is engaged in its Dynamic Systems Initiative (DSI), which is supposed to bring simplicity, automation and flexibility to the data center, while IBM, Sun and HP have already combined hardware and software in self-healing servers that are available today.

Just which vendor is leading the way in terms of a complete offering is unclear. Cisco, for instance, pretty much leads the field in terms of the networking vendors while, for hardware, IBM has certainly been blowing its self-healing trumpet for the longest as its eLiza project of yesteryear has morphed into today’s autonomic computing strategy. Elsewhere, self-healing IT forms a key part of Sun Microsystems’ N1 strategy and HP’s recently announced Adaptive Enterprise play.

Although all these vendors have certain aspects of self-healing technology already in the market, just when it will become widely available in some sort of complete offering is open to debate.

Meta Group’s programme director for infrastructure strategies, Phil Dawson, says self-healing technology will be widely available from 2004/2005 onwards, although the software stack that sits atop of it won’t be commonplace until 2006/2007.

“A network of grid-trusted computers is the delivery of self-healing concepts. However, the biggest inhibitor isn’t going to be the load balancing or virtualisation, but getting the software to fit on the model,” he says.

John Foster, volume systems manager for South Eastern EMEA at Sun Microsystems, agrees with the 2004/2005 timeline and believes the provisioning of resources, which many say are key to successful self-healing systems, will be in data centres from the first quarter of next year onwards. “The final stage is the policy automation and this will be in the data centre by 2005,” he adds.

||**||II|~||~||~|Others are less inclined to predict when self-healing IT will become the norm. One of the key reasons for this is the ever-evolving nature of technology, which means that once one piece of self-healing technology has learnt to work with all the components of a users’ infrastructure, that environment will change due to vendors/users upgrading products or introducing new ones.

“I wouldn’t put a time frame on totally self-healing networks in terms of a full end-to-end concepts because there are always new technologies emerging that have to be taken into account,” says Bassem Alkharrat, systems engineer manager at Cisco Systems Middle East & North Africa.

“With the introduction of new technologies, new applications and so on there will always be a need for new self-healing technologies. Therefore it is hard to say when you are really done with self-healing technology,” adds Miles Barel, programme director, on demand operating environment, IBM.

One thing that could potentially hold up the widespread deployment of self-healing IT is the age-old standards debate. If such technologies are to be effective in a heterogeneous computing environment, they have to be able to access other vendors’ platforms, retrieve information and then issue instructions to them regardless of who the manufacturer is.

“It is imperative that all of this [self-healing technology] is based on open and industry standards because parts of this infrastructure are going to come from different vendors. Open standards become critical so that the different parts work together,” says Barel.

With this in mind, every vendor pursuing the goal of self-healing technology is attempting to work with some sort of open standard. Microsoft, for example, is building its Dynamic Systems Initiative (DSI) around XML.

“With the introduction of XML it became much easier [for data centre components] to define themselves to each other. If everyone speaks XML, it makes it [self-healing IT] run smoothly,” says Haider Salloum, server marketing manager, South Gulf, Microsoft.

Elsewhere, IBM is using a number of standards for its self-healing solutions, including the Application Resource Monitoring (ARM) standard that helps regulate enterprise workload management, web services, and the standards developed by the Open Grid Services Architecture (OGSA) forum.

“We don’t want to create new standards where they already exist and we only want one set of standards for managing what is required,” says Barel. “The grid work, for example, has been going on for ages and one of the roles we have been doing as part of OGSA is ensuring that it doesn’t become limited to just the grid but that they [the standards] can be used to manage any infrastructure,” he explains.

Another potential pitfall to the widespread success of self-healing technology is its cost. As vendors have invested millions of dollars in developing the technology, they are unlikely to let it enter the mainstream without accruing some sort of tangible return.

||**||III|~||~||~|However, virtually all of the vendors involved suggest that the cost will not be astronomical, especially as the technology will enter the market incrementally.

“One of the things that is important under current financial conditions in the market place is that organisations need to be able to make incremental investments with immediate returns. An organisation doesn’t want to wait five years to have a self-healing system and they don’t have to because there are systems that are running today,” says Barel.

“The approach that we take allows them [users] to keep improving the levels of self management they get from their systems with simple incremental improvements to their existing infrastructure. This is important because if they have to rip out their existing technology it will be both expensive and disruptive,” he continues.

Charles Ashman, manager, ESG product business unit, HP UAE, Gulf & Levant, supports the incremental cost argument and suggests that hardware containing self-healing technology will cost the same as the previous model, if only be default.

“Because of per unit cost reductions and the commodisation of systems, we are seeing, on a per unit basis, between 10-15% reductions per year. At the same time, we are seeing [self-healing] facilities come within the products and keep the price where it is. Users are getting a lot more system and a lot more return on the dollars spent,” he says.

Furthermore, should the technology be made available for similar prices to those charged for either software or hardware products today, it means that self-healing IT will not be limited to enterprise users and those further down the stack will also be able to use it.

“With blade computing, all companies are going to have access to this technology and can run a virtualised environment for around US$25-30,000. It is not prohibitive in terms of cost and it is not limited to the big boys,” says Sun’s Foster.

Due to self-healing technologies proposed affordability and suitability for those outside of the enterprise segment, the vendors believe it will definitely take off in the Middle. “People will adopt it in the Middle East because even though they often have smaller data centre, they are still complex,” Foster says.

When self-healing technologies’ do become commonplace, users can expect a number of benefits. According to the vendors involved in such projects, these advantages include improved user service levels, reduced costs and better technology return on investment (ROI).

“The benefits users get are improved quality of service and improved reliance, both of which a company needs to be competitive and viable. Self-healing IT gives users much better ROI on IT and it allows them to take advantage of new business opportunities, provides them with accelerated time to value and gives them a competitive advantage,” says Barel.

||**||IV|~||~||~|“Some of the pain points that self-healing IT reduces on the customer side are the cost of managing the data centre environment, deploying applications and building applications,” adds Salloum.

In terms of service levels, the need for highly available IT is becoming paramount as an increasing number of organisations integrate technology with their core business process and incorporate web based technologies into their overarching corporate goals and objectives.

“CFOs and CIOs recognise the value proposition that self-healing delivers in terms of supporting business and mission critical applications and services. They [management] understand the productivity gains and the user satisfaction of self-healing technology and the zero tolerance users have today for outages in critical areas,” says Cisco’s Alkharrat.

“Very simply, companies today are relying much more on their IT infrastructure, not just as something that provides optimisations and efficiencies to their business, but as part of running their businesses… As a result, self-healing technologies are fundamental to ensuring the constant availability of systems and ensuring a business is resilient,” adds Barel.

When it comes to cutting costs, proponents of self-healing IT claim the technology does this in several ways. For instance, Cisco says that high availability equals no downtime and therefore no revenue losses. IBM and HP, on the other hand, argue that self-healing ensures servers are used to somewhere near their full capacity, something that happens infrequently at the moment.

“Wintel servers are only used up to 10% [of their capacity] and Unix servers up to 15%. Self-healing technologies can move the workload around and adjust to fluctuations, drive up utilisation and ensure that companies get a much higher ROI on the investments they have already made,” says Barel.

One method of cost cutting none of the vendors engaged in self-healing IT projects are keen to promote, however, is the fact that less human intervention means less IT staff are required to manage computing environments, thereby reducing wage bills. Instead, the vendors argue that companies can direct their human resources to value adding initiatives and focus on developing IT services, rolling out new applications and generally undertaking initiatives that sharpen a business’ competitive edge.

“Self-healing technologies allow people to have more free time to do more business orientated projects in the IT department while having the peace of mind that things are being managed correctly,” says Mohammed Amin, regional general manager of EMC Middle East.

“It is about getting them [IT staff] focused on the business. Self-healing is not about getting people removed from the equation, but about their effective impact on business,” adds Barel.||**||

Add a Comment

Your display name This field is mandatory

Your e-mail address This field is mandatory (Your e-mail address won't be published)

Security code