Application performance monitoring tools only half the battle at trucking company

Application performance monitoring tools would be fairly useless without the right people and management processes to go with them, according an IT manager at trucking company Con-way.

Improving application performance and the overall end-user experience at trucking giant Con-way Inc. meant purchasing the right monitoring tools and, more importantly, redeploying people and processes to get the most out of monitoring data.

More on application performance monitoring tools:
Application performance management tutorial: Pulling it all together  

Application performance monitoring firm targets cloud computing

Con-way, a major freight transportation and logistics firm with more than 500 worldwide locations and about $5 billion in annual revenue, had a big problem awhile back when it found that gaps between its various monitoring technologies were causing application slowness and other performance problems to go unnoticed, said Jonathan Dowell, Con-way's IT process manager. The problems resulted in frustrated end users in the field and sporadic slowdowns in the flow of freight.

"End users would complain about performance, but we weren't able to verify it," Dowell said. "Our monitoring was focused primarily within individual technology silos and those weren't correlated. We could tell you what the server was experiencing, we could tell you what the database was experiencing, but we couldn't tell you what the user was experiencing."

"The highest priority is getting the right people involved, understanding who those people are and getting them to understand their roles and responsibilities," Dowell said.

Con-way's approach to IT
Con-way's IT department is responsible for all of the software and hardware related to the company's transportation, contract logistics, supply chain management, truckload brokerage and trailer manufacturing activities.

 

The company runs a mix of Windows, Unix and Solaris operating systems primarily on Dell servers and a handful of IBM and HP servers. Those servers handle the load for client devices and personal computers around the world, which connect to each other via a Multi-Protocol Label Switching (MPLS) network and are managed from Con-way's centralized data center in Portland, Oregon.

"A large majority of those client devices connect to our Citrix Presentation server farm, where we present a [virtual] Windows Desktop to the end users," Dowell said. "We spent a lot of time migrating all of our [legacy] mainframe applications to a Java/Web-based application portal that we call the Navigator. The Navigator is hosted on some ILS (Internet Authentication Service) and Apache Web servers."

Con-way's end users log on to the Navigator via Windows Active Directory and from there they can then gain access to the business applications pertinent to their individual roles. Dowell said the Navigator hosts all of Con-way's primary business applications, including those that cover pick-up and delivery scheduling, billing and finances, human resources, and customer service.

Closing the application monitoring gap
With end users complaining of slowness and other problems, it became clear about two years ago that Con-way had a gap in its application performance monitoring strategy. The company conducted a "gap analysis" and found out, to nobody's surprise, that the biggest problem was in the area of end-user experience. Dowell and crew then set out to find a new monitoring software vendor that could improve the situation.

The team narrowed the choices down to Compuware, CA and Network General and invited those vendors in for "proof of concept" demonstrations. Dowell said CA was quickly eliminated from the running because its monitoring tools didn't work with Con-way's Citrix environment at the time. The Network General software didn't offer Con-way enough flexibility in terms of how applications are defined.

"We thought [Network General offered] a very flexible solution with an open API (application programming interface), but it would have cost a lot for us to maintain it," Dowell said. "And it didn't have the definitions [capabilities] where we could drill down and say, 'these are the transactions that are associated with this application.' We really needed that kind of granularity."

Dowell said Con-way decided to go with Compuware Vantage 10 because it was easy to deploy and didn't require the team to install agents on client devices. It also gave Con-way the ability to quickly discover, isolate and, as a result, solve the problems that affected end users.

Compuware Vantage 10 peered into individual technology silos and made it easier for the IT team to figure out if slowness experienced by end users was the result of network latency, dropped packets, slow server time, slow HTTP response, a database problem or a host of other issues, Dowell explained.

"Quickly identifying the domain from where the fault is coming from helps us to assign the right resources to investigate," he said.

Dennis Drogseth, a vice president with Enterprise Management Associates, a management software analyst and consulting firm headquartered in Boulder, Colo., said Compuware Vantage is good at helping users understand how the network affects applications and vice versa. One result of this capability is better insight into end-user experience, he said.

Drogseth added that he'd like to see Compuware continue to work on helping its customers deal with change. "Change management is one of the areas that is very hot right now, and justifiably so," he said. "Compuware should continue to make more investments in dynamically being able to assimilate and look at the impact of change, because change certainly has a big impact performance issues."

People and processes essential to application monitoring
Dowell said that getting the most out of any new monitoring technology is largely a matter of putting the right people and processes in place. For companies who are serious about improving application performance, that could mean a major reshuffling of IT personnel.

"We appointed a cross-functional governance committee that could make decisions about thresholds and about the definitions of applications," Dowell explained, "such as what they should be named, who is responsible for them, and how to deal with and prioritize whatever problems come up."

The team then assigned resources to administer the Compuware Vantage software. Dowell said those people lead the way whenever Vantage needs to be reconfigured as a result of changes in Con-way's application architecture.

Finally, Con-way appointed a process manager -- a person charged with creating and implementing formalized processes designed to optimize the new monitoring software. Dowell said the processes that Con-way built are based on a scaled-down version of the IT Infrastructure Library (ITIL) v3 that was customized specifically for dealing with monitoring technologies.

"We put a monitoring and reporting process in place, created availability and performance metrics that we report on our IT scorecard," Dowell said. "We also put together a problem investigation process, assigned a problem investigation team and ran through the problem investigation process to understand causes and implement fixes."

Tying it all together
One of the most important steps in any application performance monitoring initiative is combining data from disparate monitoring "silos" into a correlation engine and dashboard. It's an ongoing process, says Dowell, and it's something that Con-way is continuing to work on.

Dashboards -- graphical consoles where information about various systems is aggregated in a customized format -- improve overall visibility and save IT staff from memory-dependent and era-prone correlation and analysis.

"Everyone should drive towards integrating all of those tools together," Dowell said. "We're [still] doing a lot of that manually."

Con-way is also beta testing Compuware Vantage 11 and mulling possible upgrade plans. Dowell said the next version of Vantage includes a dashboard and other new resources for interfacing monitoring tools.

"It's important because our primary goals here are availability, scalability, reliability and growth at the lowest cost," Dowell said. "Monitoring [and correlation] provides visibility into those two primary things; availability and reliability."

Mark Brunelli is a contributor to SearchDataCenter.com.

Let us know what you think about the story; email XX, News Writer. Also, check out our blogs: Data Center Facilities Pro, Mainframe Propeller Head, and Server Farming.

Dig deeper on Configuration and change management tools

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchWindowsServer

SearchEnterpriseLinux

SearchServerVirtualization

SearchCloudComputing

Close