SaaS: A Virtual Data Center

William H. Matlack, Jr.
March 26, 2008 | COMMENTS

Since its introduction a few years ago, Software as a Service (SaaS) has provided small and medium businesses with affordable access to comprehensive business applications and services comparable to what much larger enterprises deploy. Because SaaS solutions only charge for access and use, businesses are spared the much higher costs of purchasing, deploying, upgrading, and managing expensive applications.

As small and medium businesses, most credit unions and their branches are good candidates for the SaaS paradigm. The size of an average credit union makes it challenging to accommodate the costs of configuring and staffing a full data center, let alone incurring the high costs of deploying and maintaining business applications.

The workload on staff can be daunting, with small teams having few resources to fall back on either after hours or in crunch times. However, because of the sensitive nature of their data and the need to ensure accessibility, many credit unions prefer not to have their processes run on external servers.

Today however, thanks to new types of SaaS offerings, many credit unions can take advantage of remote data center services, which allow them essentially to "outsource" parts of their data center services, yet run processes and systems in-house. The result is a "virtual" data center environment that's efficient and affordable for credit unions of all sizes while still satisfying the requirements for access and control.

24/7 Monitoring

Fiserv Galaxy Credit Union Solutions has provided full-service data center support for credit unions for more than 40 years. Through its online, in-house core processing, Galaxy extends data center services to credit unions nationwide. By monitoring credit union applications on a day-to-day basis, Galaxy has been able to focus on providing innovative solutions and services that greatly enhance and streamline processes that are unique to credit unions.

One such solution is a SaaS-based 24/7 monitoring solution.

One of the persistent problems and challenges for credit union information technology (IT) staff is the operation of after-hours processing. The costs and challenges associated with keeping end-of-day and end-of-period business processes functioning makes in-house data center solutions less than optimal.

Whenever a credit union needs to run these processes, there's often the need to have at least one person on-site to watch progress, identifying and solving the various problems that always seem to spring up and bring processing to a grinding halt.

It's not too bad when everything is run during the day. But day running isn't practical from either a business operations or infrastructure costs standpoint. These end-of-day and end-of-period processes of necessity run at night. This means a knowledgeable credit union IT person has to remain on the premises all night as well, which is often even true if the processes are being run at a remote data center.

"Credit unions face a real quandary when it comes to 24/7 data center operation," says Vince Francone, Galaxy's senior vice president of IT. "They need 24/7 monitoring of all their processes because they are running them day and night. However, they don't want to have to staff their IT center 24 hours a day. Having a single employee at the center just to monitor processes all night and on weekends creates problems with employee security, among others."

All-Nighters

It's also expensive for credit unions to staff data centers all day and night because if there's a problem with an application or process, the employee has to be skilled enough to be able to resolve it—and it's tough to find skilled IT professionals willing to work off hours regularly. In fact, smaller credit unions simply don't have enough IT people on staff for 24/7 operations.

Larry Reckard, manager of information systems at West Virginia Central Credit Union in Parkersburg could relate to this problem. As a one-person IT department, Reckard had to be on-site to manually run all the daily, weekly, monthly, and other transactions. And it was up to him to troubleshoot problems as they sprang up.

"It seemed like I was always on-site," says Reckard. "Transactions are the very heart of our business, so they took precedence over any and all personal aspects of my life."

The reason IT personnel need to be on-site is because of a deficiency in the capabilities of process schedulers. Schedulers are used to set up, or schedule, the many processes that need to run.

However, what often happens is that even the smallest problem with a process, like a cache that hasn't been flushed, causes the process to come to a halt until addressed by an IT professional.

The IT person usually has to spend a large amount of time determining the source of the problem. He or she then has to resolve the problem before restarting the process—which can happen several times during a single run.

So, an IT person had to sit around all night waiting for something to happen, which isn't a good use of human resources. Galaxy wanted to provide a solution for this "automation gap" in its remote data center service. The company's search led to the industry's first really intelligent job scheduler from Tidal Software.

Intelligent Automation

Tidal's approach is based on the concept of adding "intelligent automation" to its solutions. The company realized that one of the most glaring deficiencies in the market was a lack of data center automation. This was particularly apparent with the way data centers typically addressed process-stopping problems.

The typical scenario was that a problem, or several problems, would stop the processing, and IT people would gather in person or via the phone to conference and attempt to isolate the cause. The resulting effort eventually would pay off in a resolution to the problem, but it paralyzed IT resources—and both it and other problems happened over and over again.

"This inefficient scenario was really impossible for us to justify to our credit union customers," says Perry Sauter, Galaxy's technical services manager.

Recognizing that automating problem solutions was just as important as scheduling jobs, Tidal engineered a powerful problem-identification and problem-solving intelligence into its job scheduler. Beyond just setting up applications to run, the Tidal job scheduler ensures that they will run. Because the scheduler is run from Galaxy's server, it's a true "Scheduler as a Service" solution.

This intelligence allowed Galaxy to provide its credit union customers with "unattended processing." If an error stops the process—say a corrupted file or an application that's not ready to run—Tidal's scheduler flags the error and automatically brings it to the Galaxy support group's attention.

"When we see a process flag, we can intervene for the credit union by either fixing the problem or calling the credit union's IT person and asking how they want to handle it," says Sauter. "In many cases, Tidal's job scheduler may be able to automatically fix the error to continue the process without human intervention."

That way, the credit union's IT professional can go home at close of business and get a call if something goes wrong that needs intervention. And these calls become less frequent as the scheduler is used.

Faced with the expensive prospect of hiring another person for West Virginia Central's IT department, Reckard implemented Galaxy's Tidal-based scheduler.

Tidal's scheduler was engineered with two critical automation components. First, it automatically identifies the source of a given problem and notifies IT personnel. Second, it automatically fixes problems through handling rules that are set up initially and developed as the solution is used over time.

Reckard says in the past 10 months of deployment only two problems have come up. He set a handling rule for both, and things have been working perfectly. The automation of his scheduling has allowed Reckard to begin working regular hours.

Because the system runs 365 days, all customer transactions are recorded on the actual date of the transaction. Saturday transactions show up for that day—even transactions made on Christmas day show up on that day.

"The scheduler gets smarter as it is used," adds Francone. "Each time we get a flag that requires our intervention, we can set up a handling rule for that specific problem, and the scheduler will automatically solve it the next time it comes up."

Credit unions benefit from the addition of intelligent automation, like Tidal's scheduler, to SaaS solutions. It allows them to offer more and more compelling services to members while maintaining acceptable costs of doing business.

William H. Matlack, Jr. is a freelance writer and public relations consultant in the San Francisco Bay area. This story first appeared at www.creditunionmagazine.com and is reprinted with permission.