George Marootian is the Executive Vice President and Head of Technology for US Distribution at Natixis Investment Managers. This is the first article of a three-part series that will discuss the firm’s move to the cloud, continued operations in AWS, and the innovation and agility the organization has benefitted from since the migration.
When I first started as Natixis Investment Managers’ head of technology for US distribution back in late 2014, we had three priorities that guided our vision for the IT side of the business: pursue IT projects that helped the company manage risk, create or improve business efficiencies, and enable new opportunities for innovation.
We also had a more tangible and focused goal to retire our in-house data center in less than three years. This mandate was non-negotiable because the company was moving its US headquarters to a state-of-the-art, “green” space. The new building was designed to be Boston’s most sustainable office building, and we did not want to build a new data center footprint. A new data center would have been costly to build, inefficient as compared to cloud offerings, required continued capital investments, and presented potential technical challenges in a retail-oriented area.
Nearly three years later, we have eliminated our physical data center footprint (we still house a percentage of our original internal compute and storage services in a Markley Group colocation facility) and moved the majority of our workloads to the cloud. During the migration, we were also able to deliver on our mission to make IT more responsive to the needs of the business.
Engaging with CTP
To help us move to the cloud, we partnered with Cloud Technology Partners (CTP) to assist on the execution of a multi-phased project. In order to access Amazon Web Services’ (AWS) ability to deliver the cloud provider benefits we were seeking, CTP conducted a Proof of Concept (PoC). We determined that AWS was able to meet our needs and that cloud technology would provide increased speed to market, agility and improved turnaround time.
After the initial PoC, we started a full-scale migration of our production workloads and worked with CTP to ensure all aspects of the cloud’s networking, security and operational management were secured to our operating standards. Our team worked closely with CTP over six months to conduct a Total Cost of Ownership (TCO) assessment; identify all legal, compliance and technical requirements; develop the necessary account structure; establish network connectivity; build a Minimum Viable Cloud (MVC); and address monitoring and operational concerns. Then we set up a production environment that would take advantage of multiple AWS services, including VCP, EC2, S3, Glacier, CloudWatch, SQS and CloudTrail.
Choice of AWS
Choosing AWS as a public cloud provider would be a more difficult choice if we were making the decision today. Microsoft Azure and Google Cloud have emerged as compelling alternatives to AWS. Based on Microsoft’s progress as a cloud provider, a multi-could blend of AWS and Azure could be a viable approach for long-range plans.
Microsoft Azure offers services similar to AWS’ offerings along with compelling core applications in terms of services; “cloudifying” those operations might provide Microsoft with great opportunities. Google, meanwhile, has taken the concept of data science, analytics and platform computing and implemented a dynamic infrastructure over it. For firms that require large-scale data processing and analytics, Google offers a comprehensive framework to store the data and move the data around as well as carve through it to perform analysis.
That said, AWS constructed a service catalog that aligns itself with the enterprise, and we found its service offerings quite compelling. The security model met our expectations, the tools were what we wanted in our inventory and it represented a strong solid fit.
Good Fits for Cloud
As we looked into where cloud would fit best in our organization, we focused on two areas where it could provide the most value. One was generating process automation for a Natixis business division focused on overlay management. Using the cloud, we identified opportunities to create scalability, reduce processing time and incorporate automation to remove human-factor interruption. The other was with enterprise movement of data to support all business lines within our organization.
To make the first migrations smooth and timely, and to show off the power of our ecosystem, we moved the application stack for the entire division. We went through the division’s inventory and rearchitected the applications directly into the cloud. We were sure to guard against building everything so tightly coupled to Amazon that it would not be portable to another provider in the future; this ability to de-couple the open source stack that we leveraged was pivotal in our cloud strategy.
We also used the cloud to lower the risks of moving data between departments. Moving to the cloud provided transparency for our systems and identified optimal functionality. It improved the flow of portfolio and sales data and reduced the risk of delays in our daily file delivery processing. In our industry, time is an important factor so we focus on driving process efficiencies and trade execution to meet daily milestones; time-to-process metric is a key element.
We addressed some unexpected engineering challenges encountered along the way. For example, we rolled out a new framework for job automation and migrated our two-year-old services to the new environment. We will continue to evaluate and reassess our decisions and bring new services up as needed. This is a conservative approach to continuous improvement of systems, but provides far less operational disruption, which is a key performance indicator for this initiative.
Though our move to the cloud remains an ongoing process, we have succeeded not only in migrating off our data center, but also in creating a resilient infrastructure. Our systems are more stable, resilient and scalable without being bound by a particular physical location (geo-diversity is a key element of our future internal IT resiliency standards). We have reduced our risks, eliminated human-factor oversight when appropriate and saved on costs across the board. Mission accomplished.