Comprehensive coverage

Computer heal yourself

Software systems for self-management, configuration and healing and recovery
Self-trafficking will receive more and more tasks that are currently performed by
human operators. The question is how willing the system administrators will be
hand over their duties to the machine.
Karyl Scott, InformationWeek

What a shame that computers are not more like humans. when we make an effort
&#;8211 Our hearts beat faster. When we are hot &#;8211 we
sweating But in the 54 years since the British mathematician Allen introduced
Turing the idea of ​​artificial intelligence, the computer scientists failed to satisfy
Something that will come close to a self-aware computer with self-healing capabilities.

The change could be very close. Scientists in laboratories in business entities
And government officials are building systems that will automatically perform the tasks
which are currently imposed on human information technology operators. the researchers
From IBM laboratories to the US Department of Defense, capable systems are being developed
Self-adaptation that they can manage, restore and recover, protect themselves, define
Configure dynamically and adapt themselves to optimal activity
&#;8211 All without human intervention, just like our body fights infections
external without congealing thereby. Even if the final goal is years away
Next, the first generation of software ready for capable business activity
For self-recovery and independent adjustment is already with us.

Biology is more than just anology here. The researchers are looking for inspiration
In the natural world, to develop adaptive technologies flexible enough in order
To deal with the increasing complexity of computer systems and software that drive the
the world. The living cell technology or the sociology of ant communities are
Examples from the world of biology that provide new ideas for ways in them
Computer systems can focus on business goals without requiring operators
which explain to the servers how to handle each task.

"The more customers we talk to, the more we hear the same refrain,"
says Greg Burke, eLiza project manager at IBM. "The technology is complex
Too much, and computer and information technology departments are finding it difficult to meet the requirements
The maintenance of a multi-manufacturer and multi-architecture environment".

Intermediate: IBM &#;8211 Self-learning algorithms

The eLiza project team at IBM develops hardware, software and networks that will be able to
Allocate computing resources on demand, maintain data and ensure the
The continuous performance of the business activity in the event of a malfunction. results
The firsts of this development effort are already available on the market, such as software
IT management services for electronic businesses, coordinating availability
the information technology resources and among the pre-defined business requirements
&#;8211 All in order to ensure that the required performance levels are maintained

A major challenge in the implementation of self-management technology will be how to convince the
The human managers let the systems operate on their own. as in all
Another new automation technology, the transition process is expected to be difficult,
When system administrators strive to prove that they are able to bring
For a better optimization of the resources than the one proposed and carried out by me
the computer itself.

Various manufacturers are starting to offer first examples of proprietary software
Recovery capacity and self-healing. This is the case, for example, with regard to Windows, XP, the searcher
Online updates. In the past, Microsoft tried &#;8211 and with no success
Raba &#;8211 to make its systems easier to manage. the same enterprise
Management since 1997 gave rise to claims by information system managers according to which
The supposedly improved management functions cause incompatibility and failure
systems. Microsoft has been slightly better at self-healing systems
that have been integrated into its SQL server, starting in 1998. Today, it predicts that
It will take 10-20 years until new operating systems appear
and an architecture that will enable self-repair, at the level required by the business world.
"Adopting a system that will manage itself will require three steps," predicts
Craig Mundy, Chief Technology Officer at Microsoft. "First, there will be
The online automatic updates of the operating system, as it is done
In the Windows XP operating system, the next step will be established IT systems
policies, which will actually be a software add-on to the hardware and operating systems
the existence These plugins will perform automatic configurations and implementation of
Software based on the IT policy of the organization". According to Mundy, Mon
These steps will be carried out in the next three years, but will require another 10-20
Change and operating systems in completely new architectures to reach
for systems with self-healing capability. The attempts are not tools to create
Such systems are now being made in government and private laboratories.

The technology behind the eLiza project at IBM is being developed by a team
Computer scientists, physicists and mathematicians in various laboratories of the company.
Over the past decade, this team has been developing algorithms that are capable of learning
Self capable of analyzing and determining whether a computer system is working or not
which has a fault that may lead to instability. This algorithm knows
Diagnose when one or another computing resource should be reallocated to a task
certain This technology is implemented in the MVS operating systems in the environment
Central computing since 1994, and it optimizes
processes, manages the central processor configuration and other functions that have become
Too complex for manual-human management. Now, IBM is pursuing a realization challenge
These technologies are also in the world of distributed computing, and in a multiple environment
Operating Systems.

This is the place of workload management software in a heterogeneous computing environment,
developed by IBM, which relies on the company's self-learning algorithm
To automate a large part of the configuration tasks, e.g
Allocation of central processor resources currently performed by technologists
information. The software captures the system status every 10 seconds in order to monitor the
The change in performance. It is now being tested by IBM customers in the field
insurance and financial services, and its first version will be integrated into the systems
IBM's activation for servers later this year. The system is expected to offer help
Especially required: each server can be configured in about 500 different ways, and an administrator
The system needs days if not weeks to find the appropriate configuration
most to the business activity environment of the company using this server.

The longer term goal is to manage not only the servers &#;8211 but
To all components of the computing infrastructure, from operating systems and networks to
for applications and infrastructure software. "Today there is no other software that performs
Automatic configuration," says Donna Dillenberger, a senior member of the team
IBM's research. “Human operators must tune each application, on
An operating system and any interface between different parts of the system." load manager
IBM's work will also be integrated into the company's other software products, including
Tivoli Systems Management Suite.

Intermediate: A vision for automated computing
IBM's long-term vision for automated computing is even greater:
Managers will define business goals, and computers will automatically execute the
The information technology activity required to meet these goals. so,
For example, in a financial trading environment, a manager can decide which transactions are due
be perfected in less than a single second to ensure lessons
Adequate profitability. Automated tools such as IBM Workload Manager
will define the configuration of the systems required to meet these goals.

When it comes to large volumes, the effort to develop healing technologies
Myself became extremely complex. The greater the number of devices and systems
Connected to the Internet, network technology must deal with complexity
very high "We're not just talking about access from one application to the computer
personal or to a local server," says Sun's chief technology officer, Greg
Papadopoulos. "We are talking about millions of users accessing the service
via the internet. Therefore, these services and the entire network should be
More rigid than today's software. There is no alternative to infrastructure systems
that organize themselves". He concludes.

Individual components in the system may break down, but a developed genie system
Biddy Sun has the ability to realign resources and components so that the system
Everything will continue to run. Ginny is based on the idea that everything on the web is
Currently, or will soon be &#;8211 service. Whether it's software or
In hardware every service will have a discovery mechanism. When a service resource goes down,
The network heals itself by denying access to the problematic service from services
others and instruments. So far, Sun has sold 80 licenses to Genie and she
Installed in 75 sites, including the 21 bomber's self-healing system
US Navy DD.

Leave a Reply

Email will not be published. Required fields are marked *

This site uses Akismat to prevent spam messages. Click here to learn how your response data is processed.