Marz Systems, a San Francisco-based software and IT services provider, migrated a Fortune 500 biotech firm’s enterprise resource planning workloads from its on-premises data center to Oracle Cloud. The move saved the customer “more than a couple million dollars” in data center footprint expansion costs, according to Marz Systems CEO Tayyab Qader.

In an interview with SDxCentral at Oracle OpenWorld, Qader said the biotech firm was experiencing performance and scaling problems running these workloads and wanted Marz Systems to help it find a way to improve performance without expanding its data center.

“We looked into different options, and cloud was one of those,” Qader said.

Marz checked out Amazon Web Services, Microsoft Azure, and Oracle Cloud. But because the customer was already heavily invested in Oracle technology, Marz Systems went with Oracle Cloud over AWS or Azure.

Oracle’s cloud infrastructure-as-a-service (OCI) runs on bare metal servers or virtual machines.

“Although OCI is public cloud, Oracle data centers are very secure and also have all the virtual networking and virtual security layers in place, plus the flexibility to scale,” Qader said. “The best thing about the bare metal: we have better control of the server, tweaking and tuning it in order to get the best performance.”

After designing and testing the architecture, Marz migrated 12 servers from the customer’s data center to the Oracle Cloud. It took about three months because of the complexity of the systems.

Moving to the cloud achieved the biotech firm’s performance goals, Qader said. “We had almost two times the performance in the cloud compared to what they had on-prem. Reports that were running for 25 minutes [on-prem] started completing is less than 12 minutes. They went from a two-second single sign-on to under one second.”

To put this in perspective: about 40,000 employees saving one second every time they sign on to the system — Qader estimates about 15 sign-ons per day — translates into a “few million dollars” savings.

Cloud migrations also present an opportunity to perform a spring cleaning of sorts, by archiving workloads and shrinking the footprint of stored data.

Qader also recommends trying to simplify the infrastructure and to not make it more complex than necessary. “That is a lesson learned,” he said. “Also, it is better to move tightly integrated systems to the cloud at the same time to avoid performance problems.”