The aim is ambitious, the time commitment considerable and the
research and development funds immense. As the industry attempts to
remove complexity from computing, surely partnership is the best
approach?
In the presentations that have been given on self-healing autonomic
computing, more than once the expression "it would be tragic if
this was not achieved" has been heard.
What is tragic is that some of the large vendors in a position to
bring change seem set on following an independent course, rather
than pool research findings and work together.
On a trip to London last week, Fujitsu Siemens chief technology
officer Dr Joseph Reger was dismissive of the campaign IBM has been
running on autonomic computing and was keen to stress that it was
ahead of Big Blue in developing the technology.
"IBM has done a good job in marketing, but it is not the only one
doing it. We can do parts of that on our own and a lot of work has
been done which we can utilise. When we ship [the latest Blade
servers] in July, the system will have 38 functions of this type,"
he claimed.
Going it alone
For those hoping that movement towards
autonomic computing would usher in a period of widespread genuine
partnership, these words put a few nails in their coffin.
In a rush to get organic self-healing features into a product
quickly, the next generation of Blade servers - where 40 rack
mounted servers have been increased to 300 - will include fail
detection and overload self-correction functions.
"We have a six-month development advantage by coming out in July.
Dell and IBM will have similar devices out, but not until the end
of the year," Reger said.
"With thin servers, you now have a couple of hundred servers and it
is clear that individual server management is not the way to go.
Management technology that can do this on its own [is the preferred
option]," he added.
It appears that some firms will compete alone to try and solve the
puzzle of removing complexity from computing. From one point of
view, this is understandable because the rewards are significant,
but it goes against the mood in the industry that IBM has been
encouraging - for all vendors to try and get to the finishing line
together.
There are some similarities with the human genome project, which
saw the dual attempts to charge or spread the code freely divide
the scientific community. Far from working for the benefit of
mankind, the rush to get the code released was partly to prevent
the patent-happy handful of US scientists from making it impossible
to get the information for free.
One vision
The discussion on autonomic computing began
in earnest late last year. Big Blue's vision of the next stage of
computer development was encapsulated in Autonomic Computing -
IBM's perspective on the state of IT, written by senior vice
president of IBM Research, Paul Horn.
Briefly, his idea is that if computers can be designed to mirror
the human nervous system, then autonomic techniques can be built
into future computing devices, making systems look after themselves
without any human interaction.
Horn's paper sets out clearly the need for co-operation: "We call
on the entire IT industry to refocus its priorities on this
essential goal. We must co-operate in developing the necessary
standards and open interfaces to make this vision a reality," he
writes.
"This is bigger than any single IT company. It's a vision that
requires the involvement of the top minds in the technology
community," he adds.
The debate barely had time to cool before IBM was carrying on with
its vision and getting others involved in the project. In April,
the vendor held a summit to take discussions about autonomic
computing forward.
Microsoft, Hewlett-Packard and Sun joined key figures from
Stanford, Berkley, Cornell, the US government and NASA to take part
in a debate that would drive the next stage of industry
development.
Robert Morris, vice president of personal systems and storage
research and director of the Almaden lab at IBM, opened the
conference by outlining the co-operative approach Big Blue is
taking: "The purpose of the conference is to facilitate
collaboration on the science and technology of autonomic computing
throughout the research community," he said.
What price rivalry?
The amount of interest and
wide-ranging support for autonomic computing indicates that this is
a concept that is not going to go away any time soon and, despite
the differences in approach, Reger accepts its inevitability.
"We are talking about many years and once this is working for racks
and servers, the desire will come up for standalone systems. The
real savings will show if everything is included. That will take
many years," he said.
Reger argued that far from being a gimmick, the products coming out
this summer, which Fujitsu has dubbed 'organic servers', will
produce cost savings for customers and should interest the
channel.
"The advantages and return on total cost of ownership will be
visible immediately," he claimed.
Those hoping that partnership would replace competition will be
disappointed to hear such competitive talk of return on investment
and the promotion of one brand over another.
"Competition is healthy, but it comes at the end of the process
when there is something for everyone to compete on. Right now the
products coming out are nowhere near the ideal put forward by
advocates of autonomic computing, and the fear is that by fighting
over a concept which is still only half baked, these vendors will
kill it off," says one source.
"In an era when partnerships are all the rage, the failure of some
vendors to put aside their competitive instincts and genuinely
co-operate is shameful. If it is possible for a single vendor to
solve all of the industry's problems, why hasn't it happened?" he
asks.
All those taking part in the race to reach the finish line and
deliver autonomic computing accept it will take years to be
delivered. "The obstacle is complexity. Dealing with it is the
single most important challenge facing the IT industry. It is our
next grand challenge," states Horn's manifesto.
In the meantime, squabbling over who devised the term 'autonomic
computing' and who is further ahead in development is a sideshow
that threatens to relegate the debate about the future of computing
to yet another might have been in an industry littered with failed
plans.
What's an autonomic system?- A computing system needs to know itself and comprise components
that possess a system identity
- The system must be able to heal itself
- The system is always looking to get the best performance, not
settling for the status quo
- The system must be able to protect itself
- The system will anticipate the resources needed, but keep the
complexity hidden from the user
- The system must be able to configure and reconfigure itself
under varying and unpredictable conditions
IBM's position
In pleasantly jargon-free language, the
future is outlined by Paul Horn, senior vice president of IBM
Research in Autonomic Computing - IBM's perspective on the state of
information technology.
Here is his description of the task ahead: "It's time to design and
build computing systems capable of running themselves, adjusting to
varying circumstances and preparing their resources to handle most
efficiently the workloads put upon them.
"These autonomic systems must anticipate needs and allow users to
concentrate on what they want to accomplish, rather than figuring
how to rig the computing systems to get them there."