Autonomics are the future of IT

12 Comments

When I founded IPsoft in 1998, one of my main goals was to decrease the incredible amount of time IT professionals spent managing applications and tools. Over the previous decade, IT experts had become so entangled in mundane, repetitive chores, that they ended up losing the passion and creativity that drove them to the industry in the first place.

That was 14 years ago. It is much worse today. According to a recent study by performance management solutions provider BlueStripe Software, 68 percent of IT executives have invested in more than three separate application and transaction management tools, and 64 percent have invested in more than six. The result: 78 percent said their management system had become so unwieldy that they could not pinpoint where transactions slow down.

IT has entered a state of bloated chaos. Luckily, there is relief ahead. Companies like us, as well as IBM, HP and NASA, are shedding light on a new era of IT. Enter the era of autonomics.

Autonomics offer many benefits to IT departments. As the name implies, autonomic technologies eliminate or reduce human intervention required to resolve problems. Immediately after implementation, they begin to observe how engineers troubleshoot problems and execute tasks. Once the learning curve is complete, the virtual engineers become the employees.

Their adaptive nature allows them to look at each task individually and use their accumulated knowledge to execute the task in the most efficient and effective manner. And as a result, they are able to execute and oversee large amounts of routine and level one and level two IT tasks, such as database management, life-cycle application support and systems monitoring. They offer the best of both worlds, combining the work capabilities of a human with the low error rate of a computer.

But most importantly, autonomics uplift human engineers. Freed from routine and menial tasks, IT professionals can focus on creative and innovative pursuits.

It is helpful to visualize autonomic technologies as an immune system layered on top of an IT department. Most of the time, the immune system allows things to run smoothly. However, the inevitable agitator will rear its head. Like antibodies attacking a virus, autonomic systems will attempt to fix problems based on how similar situations were resolved in the past. The system will only alert the owner if all the attempted remedies have failed. While your immune system might make you aware of intruders via a runny nose, autonomic technologies usually send a more blatant signal, i.e. an error report.

Part of the beauty of autonomic technologies lies in their ability to integrate with existing service platforms. Many enterprise customers already have ticketing tools in place, and they’re not open to replacing them. Instead of replacing a company’s ticketing system, autonomic technologies can make the system run more efficiently by learning which employee is best able to resolve a specific incident and assigning the ticket to him or her.

Budgets can also be drastically reduced, because artificial intelligence can now take much off the plate of employees. We have seen some of our clients successfully automate upwards of 83 percent of routine IT tasks after implementing the technologies for one year, helping them save one-third of their IT budget. These new technologies can also keep IT departments from devolving into a tangled web of separate systems.

As IT departments become increasingly essential to successful companies, executives will soon have to make important decisions about their future. Now is the time to start looking for ways to reshape the department to perform the same tasks more quickly and efficiently. Autonomic technologies are a glimpse into an optimal future for IT, where assignments are executed by technologies, and employees have time to develop ideas that drive business.

Chetan Dube is the president and CEO of IPsoft, a provider of autonomic IT services to corporations worldwide. Prior to founding IPsoft in 1998, Chetan was an assistant professor at New York University, where he focused on deterministic finite-state computing engines. 

Image courtesy of Flickr user digitalbob8.

12 Comments

Bhaskar

For someone redirecting customer calls to the right person you might want to check out why your website UI is messed up when redirected to the Chrome browser.

Roy R.

You have clearly identified a problem but I am skeptical of Autonomics or any other automated solution. Most systems tend to be complex and by introducing another system to solve the problems of the other system, we are merely shifting the problems of the first system to the second system. Despite this being an intended effect, more often than not the second system develops its own problems which completely obscure the initial problem being so far removed from it. A system is like a baby, it won’t go away once you have it and it will grow with it’s own needs. IT departments need to find novel ways to do stuff more efficiently, but relying on new systems especially those with the promise of automation to solve the problem is simply adding bloat.

PaJc

Oddly, 12-15 years ago, just as the “large” if not necessarily bloated IT management platforms were starting to integrate/aggregate multiple types of data into what we now see today, a very basic type of “autonomic” IT engineer existed. They were usually in the form of self-running scripts, developed by IT pros to make fixing everyday problems an automatic process. But, they didn’t have the capability to “learn,” just the capability to speed through the domain(s), checking and fixing things like user profiles, .ini files, application dlls that were out of date or had somehow become unregistered. They also became identified by IT management as they first network “worms.”. Their rights throughout the domain were barely managed and their solutions were only as good as the engineer or team that wrote them. They made it possible to for a help desk team of 4 people I supervised to look like heroes to upper management by helping us maintain support and control of over 2500 desktops and 100+ servers. We dropped our help desk calls by over 62% when we turned them loose. However, suddenly we’d nearly made ourselves obsolete. I was asked to trim the team. Then, just as quickly, one mistake by one engineer updating one table that the “virtual engineer” used caused a massive failure throughout several OUs and sub-domains across 5 states. That was a bad day…

The same challenge exists today. Certainly software can be autonomic, but it is still only as accurate as the people it learns from. Even more frightening, is that complex, “learning” software may see and recognize a mistake from one incoming data point while comparing it to others, this type of program could gather a few bad data points early on, allowing a possible incorrect solution to be compounded and built upon without being noticed until once again, “that one day” when it implements a fix based upon a bad data point from long ago, causing the same enormous failure. Only this time, how does the human engineering know where to start looking in todays world of converged networks where the lines between the OSI layers are blurred.

“Autonomic” software as virtual engineers will happen. It may even become smarter and more familiar with today’s global WANs and Forests/Domains/Sites and their logical versus physical structure and data flow than their human counterparts. But who writes the software that fixes the software?

kentpawar

@PaJc

Very interesting… This should be a blog post in itself!

kentpawar

Thanks for sharing this real world application of autonomics – it made for a very interesting read.. So while autonomics can translate to cost savings and faster service ticket turn around time.. would additional resources be required to review what the intelligent agents have “learn’t” overtime and have a supervisory team that reviews the fixes before enabling the intelligent agents to perform them? Do the current autonomic s/w products allow for this to happen?

James D. Major

To conclude… IT professionals recognize that implementing industry and custom support solutions, automating services, including integrating complimentary IT solutions and services ( like change & problem management systems, asset inventory, document management, notification, service request and alert/event management systems ) is critical to achieving the efficienciencies in IT, across departments and in business services.

Brant

It is unfortunate that Mr. Dube’s company does not really employ autonomics much at all. Having been a customer of their service I can attest that they are a body shop leveraging some pretty vanilla tools. They like to spin a good story and talk about Chetan’s math background at NYU but the reality is that they have no better or worse of a methodology as the next person when it comes to monitoring events.

James D. Major

Autonomic computing is and has been available to businesses and IT departments since early on in computer development. IT specialists pride themselves on leveraging the very technologies they support to develop creative solutions that not only introduce and improve applications but also make simplify their delivery and support. True to the nature of technology, solutions and services have to evolve for businesses to stay competitive. business demands. nd simply maintaining them is not efficient

Sree

Sounds like what we initially heard as “Artificial Intelligence”, “Intelligent Agents” over the last 10 years. This sounds like old wine in new bottle

rajnish k

I am just wondering how autonomics would determine application got error. If user generate ticket then based on tickets autonomics can solve issues from predefined solutions.

Raunak

Reblogged this on Indian Entrepreneurial Ecosystem and commented:
When I was working for a IT company, I had been part of the Support and Maintenance team serving a Telecom Giant in their CRM applications. The systems of support tickets is manual and has to be assigned or accepted by individuals. It is time consuming and often missed SLAs when the log grew. Autonomics seems to take care of that.

Comments are closed.