October 11, 2013

To quickly pinpoint errors in applications before they manifest, Morgan Stanley uses real-time wire data analytics though ExtraHop, an IT operational intelligence company based in Seattle, Washington, explains Chris Kozlowski, executive director at Morgan Stanley at Interop in New York City.

Wire data is used to prioritize problems, correlate events across levels of software, alert if a problem is manifesting, mine data and determine the needed capacity for new applications or add-ons.

The problem with regular wire data is having to manually extra data from a sniffer, a network analyzer that logs traffic, which only holds a limited amount of data. Another problem with using a sniffer is trying to figure out where to deploy them in the network.

ExtraHop allows real-time wire data to be analyzed, graphed in a historical view and presented according to the device or application.

Prior to having the ability to access real-time wire data, Morgan Stanley had to perform the time consuming tasks of either going back to a specific IT staffer or perform a code sweep to solve the problem, according to Kozlowski.

According to Kozlowski, the shift from using wire data from a forensic science to a solution that predicts application behavior is more efficient. Real-time wire data also provides an explanation to how an application behaves and the ability to mine useful information.

Aside from using wire data to figure out IT pain points, Morgan Stanley also measures how much a web portal, as well as the applications within the portal, are being used. One recent example was trying to estimate how much code would need to be changed for a rebranding effort.

What Do Resolution Plans Say About Data Governance (Part 1)?

“We recently had a branding change that was being considered for a domain. It was exactly one application with hundreds of distinct applications behind it. Before we can change the domain name, we would have to ask ourselves how does this application use cookies? Cookies and domains are kind of intertwined,” states Kozlowski.

Although figuring how the application uses cookies is not an out-of-the-box feature, ExtraHop is an extensible solution, so Morgan Stanley was able add to commands the product to track the cookies and quickly make a list of the value of each cookie.

According to Kozlowski, using wire data is an unconventional method for measuring how the application uses cookies. However, having wire data in real-time allows the easier and quicker data mining then conventional methods.

To extract the information off the wire, Morgan Stanley wrote less then a dozen lines of Java, whereas the alternative would have been to ask developers for the data or do a code sweep, which are time-consuming methods, adds Kozlowski.

Zarna Patel is a staff writer for InformationWeek's Financial Services brands, which include Bank Systems & Technology, Insurance & Technology and Wall Street & Technology. She received ...