An Architected Journey to the Cloud

The Risk

Our client, a major provider of fixed income portfolio analytics had operational issues with its current environment including:

• Inability to keep software versions current
• Network security issues due to ad hoc network design and firewall rules
• Deployment issues due to differences between production and development environments
• Excessive cloud service provider costs
• Inefficient licensing of expensive products due to system topology

The Solution

The OpStack solution reengineered and implemented their cloud environment based on a fully software-defined infrastructure (everything built from code) on a major public cloud service to:

• Deploy a properly segmented network controlled by Palo Alto firewalls
• Build all new Windows and Linux servers from cloud provider base images and automatically apply patches to all layered software
• Migrate existing VMs running bespoke application code using cloud provider Site Recovery
• Move Oracle databases across platforms without requiring an upgrade to Enterprise Edition — lowering ongoing licensing costs — an immediate six-figure savings
• Provide a same-day BCP recovery capability through replicated storage and automated re-creation of the server environment to a remote separate availability zone
• Provide full monitoring of both operations and security to an MSP’s managed NOC and managed SOC
• Reduce licensing costs by five figures per server through right-sized discrete Oracle instances for both production and non-production database servers
• As with all OpStack engagements, the Stack is fully documented in a knowledge management system that was transferred to the client as part of the turnover to production

The Result

• In the first 90-days of the engagement OpStack moved 100% of its infrastructure off of it's expensive legacy environment and reduced software licensing costs — savings that paid for the implementation in less than a year.
• All open infrastructure security vulnerabilities have been closed.
• There have been no production outages on the new infrastructure.
• Monthly batch processing of large data feeds into the analytical system has been sped up by a factor of four.
• The client was able to conduct its first fully successful DR test, recreating its full infrastructure using the stack’s automation scripts in a remote region within two hours.