One of the most trying aspects of technology is troubleshooting. As Sherlock Holmes said, once all other possibilities have been removed whatever remains, however unlikely, is the solution. Systems administrators enjoy solving problems but they tend not to be too keen on having to rely on Homesian logic to do so.
Or-Yehuda, Israel-based Mutek Solutions Ltd. has taken a chapter out of the aviation industry’s book and created a black box solution similar in methodology to those found on today’s commercial airplanes. The company’s Black Box Flight Recorder tracks and records the on-line operation of chosen applications, delivering system administrators alerts should problems arise and pinpointing fail points should crashes occur. No more searching for needles in haystacks, the company claims.
“Others can tell you there is a problem…but the one thing Mutek can tell you is why,” said Boston-based Glen Hartman, vice-president of marketing.
“We are about the why,” he said.
By quickly identifying the exact location of the failure, time is saved.
Mary Sullivan, team manager at Thompson Financial in Boston, which has been using the Mutek solution for about six months, said prior to using the Black Box help desk members would have to dial into a client’s site or have them send their database to Thompson Financial.
“We have had cases where we could resolve client issues probably two or three times faster than we normally would have,” she explained.
the three levels
The Black Box Recorder can be installed on a wide array of hardware, from laptops to servers, and can gather data from three different levels. The Recorder is installed on individual machines and the results are sent to a Black Box Manager.
In the case of a server farm, the manager would be installed on one server and the recorder on the remaining servers.
At the user level it can record everything from keystrokes to mouse movements to clicks and screen operations. The resulting report can be played back as a movie of the user’s experience or as a detailed text document, noting all actions taken.
Applications can be tracked at the system level observing such things as registry access, network connections, file systems and threads.
“Say a person is trying to access a Web server and a select statement fails,” Hartman said. “The system administrator can go in live and monitor the situation, even though the server or app may not have crashed, and can replay everything that the user did to see where the problem is.”
The alerts can be configured to identify specific actions or respond dynamically to any kind of runtime event, said John Schulte, director of information technology at Mutek’s Raleigh, N.C., office. The system administrator sets the profile to record what he or she wants, he added.
The third recording can occur at the code level, which is a nice offering for developers tired of trying to find snafus.
At the code level actual internal operations of programs are recorded.
“The thing that is really different from…the other people is the depth of the information that they can collect without you having to instrument the application,” said Jasmine Noel, director, systems and applications management with the Hurwitz Group in Framingham, Mass.
“What we are really allowing them to do is look inside the guts of the application and see what is going on,” Schulte said.
The product has a fairly small footprint but still gives users line by line detail if it is needed, Noel added.
“I think that anybody who develops software should be interested in this.”
For now the solution is available for the Windows platform but will be available to work on Sun’s Solaris later this year.
Cost for a typical entry level package (www.mutek.com) runs around US$60,000 to US$70,000.
Mutek, in Raleigh, N.C., can be reached at (919) 878-3717.