Davis + Henderson, founded in 1875 as a cheque printing firm, now provides back-end software and services for Canadian financial institutions. Most of its internal software is developed in-house and not purchased off-the-shelf.
With 200 servers and about 1,000 workstations across its seven Canadian locations, managing the systems — detecting potential problems and upgrading hardware and software — is no small task.
“It becomes extremely difficult to keep your eye on all those things without some kind of electronic eyes,” said Mark Bryant, Toronto-based Davis + Henderson’s director of technology operations.
This electronic set of eyes is just one aspect of systems management, which encompasses a wide variety of software and services, including patch management, asset management and alerting tools that help keep servers, networks, enterprise applications and workstations running properly.
While software agents that track specific servers, network hardware and other devices have been around for decades, several vendors have recently developed more sophisticated systems that can trace the root cause of specific problems and predict the effect an outage or slowdown can have on specific business services. Davis + Henderson uses four of the Unicenter-brand software tools, manufactured by Islandia, N.Y.-based CA Inc., formerly known as Computer Associates International.
Get back to your roots
The company uses CA’s Asset Management, which includes automated discovery, hardware and software inventory, software usage monitoring and licence management. It also uses CA Distributed Software Management and Service Desk, plus Network and Systems Management (NSM). NSM is designed to identify and predict problems as well as help IT managers decide what takes the highest priority according to its potential impact on business.
“It allows my guys to action something before it becomes a problem,” Bryant said, adding Unicenter helps him monitor the use of memory, central processing units and disks. “If it fails, we can write a script to have it restart.”
When an IT-driven service fails, determining the root cause is a “significant challenge,” said Brant Hanbury, IBM Canada Inc.’s Tivoli national brand manager. Business Service Manager, which is part of the Tivoli IT Service Management product line, includes auto-discovery, a Web console and a graphical view of applications. It is targeted at executives who want a “dashboard” view of their systems and is designed for non-IT business managers, Hanbury said.
The Tivoli service level advisor is “essentially the same thing” as BSM, Hanbury said, but it provides a different view of the same information. Service level advisor lets IT managers define service level agreements with a wizard, and can evaluate them as often as every hour. It is designed to let IT departments fix problems before they occur. “
Many clients look to end-user monitoring tools to give them an indication of why clients are not receiving adequate service,” said Cameron Haight, research vice-president for the Gartner Group Inc., a Stamford, Conn.-based market research firm. “They are the canary in the coal mine.”
Systems management tools today are “more tightly coupled to the services they’re trying to deliver,” than they were in the past, said Darin Stahl, lead analyst at London, Ont.-based Info-Tech Research Group.
Seek the single pane of glass
Over time, IT managers tend to buy a variety of tool sets from different vendors, all which give information on different parts of the IT infrastructure, but not with a bird’s-eye view. The holy grail of systems management is a “single pane of glass” that will give users an overall view of their system from end to end, Stahl said.
“If (mobile devices) never get plugged into a blue cable into my router, how do I discover those and how to they get patched?”Darin Stahl, analyst>TextAsset management is another important part of systems management, and it goes beyond keeping a database with the product name and serial number, Stahl said. For example, automated discovery is important, especially when some users are installing products without telling IT, and wireless devices hooked up to the network are not always obvious. “If (mobile devices) never get plugged into a blue cable into my router, how do I discover those and how to they get patched?” Stahl said.
Track the transaction action
Another important development in systems management, according to Haight, is the emergence of tools such as CoreFirst from vendors like New York-based
OpTier Ltd. that track specific transactions throughout the IT infrastructure. CoreFirst is designed to create a visual map of transactional relationships between servers and is designed to predict problems with scalability.
The transactions it tracks can deal with a range of issues including securities trades, requests for quotes, or attempts by users to log on to their networks, said Yori Lavi, OpTier’s chief product officer.
Lavi said one of his customers was using IBM’s messaging queuing (MQ) technology to transmit up to 30 million transactions per day between a BEA WebLogic system and an Apache HTT Server. WebLogic was slow and hogging too much memory, but when the administrators examined the root cause, they found it was the MQ that was lagging. If they had not analyzed the flow of the transactions, they wouldn’t have identified the MQ factor, Lavi said.
Get to the business context
Monitoring system use in the “business context” is a major trend, said Tom Bishop, CTO for Houston-based BMC Software Inc. “Today the vast majority of organizations — 75 to 80 per cent — have a pretty difficult time identifying the business impact when a problem occurs,” he said.
Another major challenge is keeping track of all the changes IT staff have made and predicting the effect they could have on other parts of the IT infrastructure.
“We’re not very good at making the changes in an appropriate way and understanding the impact of making the change,” he said.
Recording changes is a major priority at Davis + Henderson, Bryant said. Using CA’s Service Desk, he can find out who made a change, who approved the change and when they did it. The changes can also be made without disrupting users as one of their recent software upgrades was done after hours. “We put this stuff out while people were sleeping,” Bryant said. “It was a considerable savings in hours and overtime.”