For much of its 25-year history, catastrophe modeling and risk management firm RMS delivered its software using an on-premises model. The company's modeling software is designed to help customers mitigate financial risk in advance of events such as earthquakes, hurricanes and even terrorist attacks.
Four years ago, RMS "started down a journey of taking our solution to the cloud," says Paris Georgallis, senior vice president of cloud platform operations. RMS started with Amazon Web Services, “which provided us with the runway we needed to get our development off the ground and to rewrite our product for a cloud platform."
Over time, though, RMS decided to bring its software back in house. Driving the decision: Security, predictable performance and economics. The firm worked with EMC and cloud computing consultant Kovarus to build its private cloud.
Cloud Environment Mixes Unified Computing, Virtual Machines
RMS selected Cisco Unified Computing System (UCS) as the compute platform, VMware VSphere for the virtualization platform and four EMC VMAX 40s for storage. The firm also uses Infoblox and Chef scripting to automate the creation of VM’s, deploy the operating systems, and apply final configurations, Georgallis says.
The VMAX platform was selected after testing a range of storage options under full-load conditions, including all-flash arrays. It provides a number of benefits, he says, including automated storage tiering. "It's a 'set it and forget it' box that we have come to rely on."
[ Related: EMC Touts VMAX Cloud Edition Storage Offering ]
Colin Gallagher, senior product marketing director for EMC, says products such as VMAX can bring together the reliability of the traditional data center and agility of the cloud. This lets customers become their own internal service providers. As more customers build "mission-critical clouds," he adds, they say it's important to match cloud infrastructure to workloads, specifically when it comes to performance, availability and security.
RMS deployed this solution in four locations: One development lab, two production sites and one disaster recovery location. It's now the primary facility for delivering the firm's revenue-generating products and services, whether SaaS-based or hosted-managed.
Better Performance, Lower Costs
Private cloud in place, RMS launched the cloud version of its RMS(one) risk modeling and analytics product. Immediately, the company noticed significant improvements in performance, Georgallis says – and the cost of the environment compared to the public cloud approach has decreased by 50 percent. In fact, RMS has been running at full speed, under a full load, for over a year, and hasn't experienced any of the interruptions generally associated with public cloud platforms.
RMS can also readily address customers' security concerns. For example, Georgallis says, auditors can identify which racks, and even which servers, house a particular customer's data.
Georgallis admits that building a cloud solution is fraught with challenges. Choose your partners and solutions carefully, he advises. Pick vendors that support open source, as they will support the growing number of organizations entering the technology marketplace.
Plan for a constantly changing environment, too. This ensures that you can leverage all investments as new options enter the technology market. In the case of RMS, Georgallis says, this has meant evaluating bare-metal options and, as a result, equipping its UCS platform with internal USB boot drives.
This story, "How RMS put its catastrophe modeling software in the cloud" was originally published by CIO.