Raymond James
Raymond James Financial, Inc. (NYSE: RJF) is a leading diversified financial services company providing private client group, capital markets, asset management, banking and other services to individuals, corporations and municipalities. The company has 7,100 financial advisors serving approximately 3 million client accounts in more than 2,900 locations throughout the United States, Canada and overseas. Total client assets are $637 billion. Public since 1983, the firm has been listed on the New York Stock Exchange since 1986 under the symbol RJF. Additional information is available at www.raymondjames.com.
Business Need
Raymond James Financial conducts business with its affiliated financial advisors through their web-based technology platform, Advisor Access. The platform, which allows advisors to transact financial business on behalf of investor clients, resides behind the corporate firewall. Advisors access the platform via a local area network (LAN) or externally through a virtual private network (VPN).
“We execute all the financial transactions submitted by our affiliated advisors through Advisor Access, so they are mission critical for the health of the business,” emphasizes Jeff Palmiero, an Application Performance Management Architect with Raymond James Financial. “If the platform experiences problems, our brokers can’t conduct business, and we risk our ability to maintain long-term relationships.”
When Palmiero started working for Raymond James, he recognized the importance of monitoring the application programming interfaces (APIs) with which advisors interact. The application testing team already leveraged an effective tool for testing the application presentation tier but needed an additional tool for testing the services tier that interacts with the presentation tier.
Specific capabilities Palmiero sought included testing graphical user interface (GUI) services 24x7 and ‘warming up’ servers after performing maintenance. “If financial advisors access a server right after a maintenance session, they could experience delays as the applications on the server need to warm up at first,” Palmiero explains. “We wanted to leverage a tool to synthetically access the applications right after a maintenance session so the first real users would have immediate access.”
Palmiero also hoped to find a solution that would allow the definition of maintenance windows so system admins would not receive down-application alerts during blackout maintenance windows. The team also required a testing tool that could negotiate various authentication schemes when communicating with web services applications.
Solution
Raymond James considered several solutions for testing the performance of its advisor-facing applications. “We eliminated many of them early in the process because they did not offer an appliance for monitoring web services applications like ours that operate behind our firewall,” Palmiero says.
The technology provider that did offer this critical capability was SmartBear, with its AlertSite for API Monitoring solution. The technology monitors APIs and ensures critical application functionality performs as expected by integrating two SmartBear tools— SoapUI for pre-deployment testing and AlertSite, a private monitoring appliance to measure post-deployment web performance inside the firewall.
“AlertSite ensures the quality of our APIs, and we particularly benefit from the ability to share test scripts across each team,” Palmiero said. “This streamlines our testing processes, and when we start monitoring application performance, we can leverage the same scripts used to test the applications prior to going live.”
Raymond James application developers and the QA team leverage AlertSite by building SoapUI scripts as they conduct application and integration testing. They also use SoapUI to run web services SOAP and RESTful transaction requests, and the artifacts they create can then be promoted with minimal work into synthetic transactions.
“This gives us a base foundation for running tests 24x7 through the AlertSite InSite appliance once the applications behind the firewall go live,” Palmiero says. “The synergy between SoapUI and the AlertSite appliance is a huge technical benefit.”
Another key feature offered by AlertSite is the ability for Raymond James to execute synthetic transactions to ‘warm up’ applications when recycling the infrastructure supporting JBoss or a patched Microsoft application. Because the synthetic transactions are the first to access such applications, Raymond James does not have to worry about end users encountering issues. The application will be warmed up ahead of time.
This feature proved helpful when Raymond James launched is first native mobile application from the Internet, and the user devices required mobile device management software to access the application. When the application was deployed to mobile devices, it authenticated itself to the web services using two-way SSL.
“We used AlertSite to emulate the two-way SSL so we could make the same calls as those made from the mobile devices,” Palmiero points out. “Through the synthetic transactions, we knew if the front door closed or if transactions failed because of the two-way SSL. This was key because that would have impacted many end users.”
With AlertSite tracking application performance 24x7, we have the ability to resolve issues well before they impact our business partners and customers. When monitoring application services, we can also record transactions that exercise our load balancer as well as the individual nodes of our server farm. If there is a performance issue, the system immediately alerts application admins as to which nodes experience the issue
— Jeff Palmiero, Application Performance Management Architect
Benefits and Results
AlertSite for API Monitoring provides Raymond James with a single view into the performance of all financial-advisor applications running behind the firewall as well as an external application that individual investors access via the Internet. “Ninety percent of what we monitor involves applications accessed by advisors on our internal LAN,” Palmiero says. “But we also interact with the customers of our advisors via a public website, so it helps to have a solution to monitor both types from a single console.”
With the applications running on the internal LAN protected by a variety of authentication schemes, transactions recorded in the SoapUI component of AlertSite can negotiate those schemes and present the necessary credentials to execute the tests. AlertSite’s network of more than 80 global monitoring locations also enables Internet-facing applications to be measured from a majority of ISPs in the same way customers interact with the applications.
“With AlertSite tracking application performance 24x7, we have the ability to resolve issues well before they impact our business partners and customers,” says Palmiero. “When monitoring application services, we can also record transactions that exercise our load balancer as well as the individual nodes of our server farm. If there is a performance issue, the system immediately alerts application admins as to which nodes experience the issue.”
Another benefit Palmiero underscores is that all parties involved in analyzing tests receive an accurate reflection of the response time of GUIs and web services. This feature is particularly helpful because users often only consider the time it takes to load the part of a website page they can see above the fold. Users may not consider the time it takes to load the entire page. Knowing the time to load the entire page is critical because long load times drain server resources.
Raymond James makes sure scripts function correctly by running them through SoapUI prior to deploying them to the AlertSite appliance. “If we're going to call a synthetic transaction, then we've got to make sure that that service account or that human is added to the Active Directory group used to authorize that particular service,” Palmiero says.
When looking at a view of synthetic transactions run against the client center or the infrastructure, Raymond James can see if SOA services and the micro services that end users depend on are running properly. If micro services fail, then the client center will fail.
“That means we actually need to be able to see both, and that’s what we get with AlertSite,” emphasizes Palmiero. “We actually like it when red bubbles to the top. If you look at some of these groups, for which and the number of synthetic transactions number in the hundreds, there's no way to really show 100 things on a screen. So we appreciate that AlertSite simply indicates if everything is OK of if anything is currently erring out. We can then bubble that to the top so we can address it.”
“The ultimate benefit that AlertSite delivers is giving us the ability to create synthetic end-user sessions that interact with our applications in exactly the same way as our financial advisors and their customers,” Palmiero concludes. “We can conduct tests before applications go live to ensure they are ready to transact business and then monitor them continuously in case any future conditions impact their performance. With AlertSite tracking application performance 24x7, we are able to resolve issues well before they impact our business partners and customers.”
Business Challenges
- Ensure execution of partner financial transactions.
- Maintain partner relationships.
- Test web services application performance behind the corporate firewall.
- Streamline application testing processes.
SmartBear AlertSite for API Monitoring Solution
- Monitors speed and availability of service APIs, web services and RESTful services.
- Enables developers, testers and IT to collaborate more easily.
- Ensures API quality and scalability.
Measurable Results
- Identifies application issues immediately.
- Tracks GUI and web services performance.
- Warms up applications with synthetic transaction after maintenance sessions.
- Prevents false application performance alerts.
- Interacts with advanced authentication to enable testing of all user scenarios.