Skip to main content
Industry

RMS: Risk manager reduces its own risks by protecting data in the cloud

RMS thinks about the unthinkable. A leading catastrophic risk modeling company, RMS helps insurance companies, capital markets, corporations, and governments think through what would happen in the event of earthquakes, hurricanes, floods, terrorist attacks, infectious disease outbreaks, and other calamities. The goal is to help these organizations better prepare for and protect their constituents and assets.

RMS makes software that simulates such catastrophes, and customers can run this software in their own datacenters or in RMS datacenters as a hosted offering. RMS has multiple datacenters around the world, which are virtualized using VMware software. Several datacenters are dedicated to disaster recovery (DR), to disaster-proof the delivery of RMS risk modeling services.

“We have to think about the unthinkable in our own business, which is the possibility that one of our datacenters might experience a catastrophe,” says Pierre-Jean Olivon, Infrastructure Manager at RMS. RMS had an entire datacenter dedicated to defending against this possibility, and it used VMware Site Recovery Manager as its DR orchestration software.

However, RMS found VMware Site Recovery Manager to be too complicated to enable frequent DR tests, which are necessary to ensure that everything is working properly. “When your test process requires some 200 steps, you just don’t want to do it,” Olivon says. “We would have to pull 15 people off their jobs to perform a DR failover test and spend a good month in advance preparing. It was really hard on the business, so we avoided it.”

Additionally, RMS wants to focus on creating the world’s best risk modeling software, not running datacenters. “There’s a lot of expense involved in running datacenters,” says Olivon. “There is constant hardware provisioning, refreshing, and capacity management, as well as software licensing and support costs. We wanted to let someone run that part of our business.”

Move DR into the cloud

RMS decided that someone should be Microsoft. “Azure and Amazon Web Services were pretty similar feature-wise, but we were already a Microsoft Enterprise Agreement customer so our pricing was much better with Azure,” Olivon says. “Also, AWS didn’t have a good DR product to compete with Azure Site Recovery.”

Azure Site Recovery is a cloud-based disaster recovery orchestration service that coordinates the replication and recovery of Hyper-V and VMware public clouds across sites. It also provides nondisruptive testing of recovery plans and remote monitoring of cloud health. Organizations can protect applications to their own second site, a hoster’s site, or to Azure.

RMS is currently using Azure Site Recovery to protect on-premises workloads to Azure but eventually will move production workloads into Azure, too, and employ the product’s Azure-to-Azure DR capability. RMS uses Azure infrastructure services—Azure Virtual Machines, Azure Virtual Network, and Azure Storage—to host its DR environment.

Better protect business-critical workloads

By using Azure Site Recovery, RMS feels that its business is better protected, because its staff can easily test its DR environments for customers. “We previously tested DR only when customers requested it, once or twice a year,” Olivon says. “With Azure Site Recovery, DR testing is easy—just one click—so we can do it whenever customers want. Everything’s automated. It makes us way more comfortable with our DR setup.”

The test and recovery process is also a lot more resource-friendly. “With Azure Site Recovery, the recovery process requires far fewer people. The VMware solution required 15 people to perform a DR failover; Azure Site Recovery requires three. We used to spend a month preparing for a DR test; we now spend two weeks,” says Vineel Palla, Cloud Systems Engineer at RMS. “When you have a small team, it’s very hard to redirect 15 people from their normal jobs to a DR test, which doesn’t add value to the business. We now have more resources to focus on onboarding new customers and delivering products on time.”

Reduce costs, innovate faster

RMS has decommissioned its dedicated DR datacenters, a multimillion-dollar expense. “Whatever I save on DR I can put into our development team to innovate faster, improve product performance, and be more competitive,” Olivon says. “We have more funds to focus on improving our application.”

In addition to using Azure for DR, RMS is using Azure infrastructure services to save time and money on dev/test. “By using Azure for dev/test, our dev teams can self-service their infrastructure needs with easy setup policies, and we get predictable costs and pay for the resources by the hour,” Palla says. “By comparison, it can take a month to get dev/test hardware provisioned on-premises, and that’s assuming you have the datacenter space.”

For day-to-day dev/test server requests, server provisioning times have dropped from two days to two hours. For larger project-based server requests—getting a larger infrastructure set up in Azure—infrastructure provisioning has dropped from two months to 10 days.

“With Azure, we can worry less about capacity management; for example, we don’t need to know how much storage we’ll need next month,” Olivon says. “It lets us focus on automation and on delivering product faster.”

In fact, RMS developed RMS(one), its next-gen risk modeling solution, in Azure and offers it exclusively as a cloud-based offering.

Support global growth

RMS already has a datacenter footprint in the United States and Europe, but Azure will help it expand into other parts of the world. “Azure gives us more business expansion options, because we no longer have the huge hurdle of setting up datacenters,” Olivon says. “Also, having Microsoft as a partner is a big asset. Many customers already have workloads in Azure, and people trust Microsoft. There’s real value in our partnership with Microsoft. It’s something we advertise.”

Read more on the Microsoft Banking & Capital Markets and Insurance blogs.

Follow us @msftfinserv