A 50-terabyte database migration cost Alinma Bank 12 hours of downtime, while backing up and restoring a new policy administration system took Penn National Insurance nearly 16 hours. Both organisations faced the same constraint: Moving to the cloud was too expensive, but building and managing everything on premises required too much work.
During a panel discussion as part of Nutanix .NEXT 2026 in Chicago, Musab Umair Malik, Senior Database Consultant at Alinma Bank, and Kurt Richards, Database Administrator at Penn National Insurance, explained how Nutanix helped them reduce database latency, migration time, and downtime.
Time constraints
At Alinma Bank, non-standard deployments, storage infrastructure latency, and human error prevented the database environment from scaling. With around 700 databases and only a couple of clusters, expansion and decentralisation were impractical because of the complexity involved and the limited time available.
Although the system was properly configured, the bank experienced average disk latency of 1,000 milliseconds.
Alinma Bank eventually migrated to new infrastructure. However, given the scale of the environment, including one database of around 50 terabytes, the move resulted in 12 hours of downtime.
At Penn National Insurance, traditional backup restores took too long. When the insurer introduced a new policy administration application that needed to be backed up and restored in a test environment, the backup alone took nearly eight hours, followed by another eight hours for restoration.
Making the move
With that level of downtime no longer acceptable, both organisations adopted Nutanix Database Service (NDB), a database-as-a-service platform for provisioning, operating, and scaling databases across on-premises and hybrid cloud environments.
The platform supports SQL, NoSQL, and vector databases, including Oracle, Microsoft SQL Server, MySQL, MariaDB, PostgreSQL with pgvector, EDB, and MongoDB. It provides lifecycle management functions including provisioning, patching, cloning, backup and recovery, high availability, and security through a single management interface or REST APIs.
“When we migrated to NDB, it took less than an hour,” Malik said. After the bank moved its full workload to NDB, disk latency fell to 0.4 milliseconds.
“We’ve been using NDB for almost a year. Before that, we patched SQL Server a couple of times every quarter. We had around 30 minutes of downtime, which was acceptable for the activity we wanted to perform. Since moving the applications to NDB, however, we haven’t had any downtime,” Malik said
Technically, patching causes an interruption of five or 10 seconds, which the bank does not classify as downtime.
“NDB handles all the patching. It succeeds 90% of the time, and even when the remaining 10% is unsuccessful, there’s no downtime,” he added.
Because Alinma Bank’s batching activities took quite a while, the IT team started investigating the root cause of the slowness of the systems — but that also took hours, even days.
Alinma Bank’s batch processing activities also took considerable time. Investigating the cause of system slowdowns could take the IT team hours or even days.
“After moving to Nutanix, uptime increased from 98% to 99.999%,” Malik said.
For Penn National Insurance, NDB provided several of the database functions the insurer needed.
“We examined NDB’s snapshot function and its Time Machine capability. The multi-terabyte database was snapshotted in a few minutes and then cloned or restored. The process now takes only a few minutes, compared with 12 hours previously,” Richards said.
The insurer also preferred Nutanix’s hyperconverged infrastructure, in which storage was not managed as a separate product.
“That caused problems for us in previous years. We would have databases running and then run out of space. I would have to contact someone on the infrastructure team who understood the storage infrastructure. If it happened at three in the morning, I had to hope I could reach them, wake them up, and make the change. That process has now changed,” he explained.
‘If I’m running out of space, the system alerts me that capacity is low. I can then go in and expand the storage myself,’ Richards added.
According to Ashish Mohindroo, Senior Vice President and General Manager of Nutanix Database Service, customers increasingly want flexibility and choice in how they manage their databases.
“They want to choose their infrastructure, database, operating system, virtualisation environment, or container environment, while retaining control. In some regulated industries, that control is mandatory because data residency laws can prevent organisations from moving data to the cloud,” he pointed out.
Mohindroo said NDB allows customers such as Alinma Bank and Penn National Insurance to use cloud-style database management while retaining control over where their databases run.
“They can automate database management in their own environments, whether in on-premises data centres or the public cloud, and achieve the associated productivity gains and cost savings,” he said.
Technical support
Alinma Bank had to move around 700 Microsoft SQL Server databases from its old infrastructure within six months. Malik said Nutanix’s professional services team assisted with that process.
“Those 700 databases contained around 50 terabytes of data. The professional services engineers helped us analyse how to design the new systems and create PowerShell and SQL scripts. Although NDB provisioned the database instances, we still had to migrate them manually, which was where the professional services team helped us,” he said.
Malik also said the bank plans to expand its use of PostgreSQL, although most of its workloads will remain on Microsoft SQL Server.
“We had around four virtual machines in the old infrastructure. We currently have 88 Zero virtual machines, or Zero-VMs. Previously, we had around one Windows cluster covering the entire environment, including production and disaster recovery. We now have around 100 virtual machines across around 25 Windows clusters. We decentralised the systems from what was effectively a single Windows Server failover cluster to around 25 clusters,” he said.
Penn National Insurance has rebuilt most of its infrastructure. Some test servers remain in the old environment, but these will also be rebuilt as a greenfield deployment.
“Nutanix’s help and support assisted us throughout this process,” Richards said.
