LLNL is changing the way much of its data and applications are managed and stored. The Livermore Information Technology (LivIT) program, which services the Laboratory’s IT needs, has been actively pursuing its goal to be cloud-centric. In fact, LivIT is the first organization at LLNL to commit to migrating all services and applications to the Amazon Web Services (AWS) cloud.
With this migration, LivIT is leading the Lab’s movement to more modern and secure IT solutions that transform digital workspaces and enhance cybersecurity. AWS is a platform that allows services, servers, and applications to be stored and maintained virtually, instead of in the Lab’s on-site data center where data and back-ups have historically lived. The cloud also offers machine learning, analytics, and other innovative capabilities that can be utilized when needed. The implementation of AWS at Livermore is called ‘LivCloud.’
“Benefits of the cloud include all of the ‘-ities,’ that is, the security, scalability, agility, reliability, and flexibility of future computing,” says Judd Sproba, LivIT’s AWS Platform and Operations team lead. These benefits enable increased efficiency, greater storage capacities, and faster provisioning and deprovisioning of applications, all while strengthening cybersecurity, ensuring the safety and compliance of the platform, and protecting the Laboratory’s assets.
From a user perspective, not much will change, as most applications are web-driven and can still be accessed from their original websites. The major differences will be seen and felt behind the scenes on the IT side, as LivCloud will alleviate the need for server procurement and sizing, service management, and operating system maintenance. This move will positively impact LivIT, as system and application maintenance are extremely time-consuming. In addition, the built-in redundancy of the cloud will eventually result in higher availability and less down time for applications.
The cloud is a strategic focus for LivIT and involves many key LivIT staff. The project is led by Deputy Chief Information Officer for Cloud Transformation, Robyne Teslich. Leading up to the migration, LivIT’s Chief Technology Officer, Greg Herweg, spent several years investigating and preparing LivIT for the move. Sproba works in the data center managing migration requests across the board, while Troy Weaver handles specific applications. Other members of the 50-person AWS migration team include database experts, subject matter experts, project owners, team leads, managerial staff, AWS consultants, and senior management.
Prior to initiating the full migration, LivIT performed trial runs on specific applications, which were a huge success. For example, since its move to the cloud, Splunk, an aggregator responsible for storing computer logs across the Lab, can now handle four times more data without errors and at a reduced cost. Continued success of the LivCloud migration effort could even lead to the Laboratory acting as a consulting liaison to other companies considering a similar move.
Planning and preparation for the migration to LivCloud began in October 2020, and the team hopes to complete the migration in January 2022. The team’s goal is to have all LivIT services utilizing AWS. With hundreds of servers and applications requiring migration, the work is rigorous. To date, LivIT has successfully moved several major technology stacks and applications, consisting of over 200 applications and over 100 servers.
The team enlists the help of two special applications during the migration process—Flexera and CloudEndure. Flexera analyzes data, maps applications and costs, and tells the LivIT team how to efficiently move the applications to the cloud. Next, CloudEndure performs what is called a “lift and shift,” in which the data is prepared for the move, lifted, and shifted directly into the cloud. CloudEndure installs agents that provide a block-level, rather than file-level, copy of systems and applications.
“Once the preparation process is complete, the migration of on-site data to the cloud can occur with a ‘flick of a switch,’” says Weaver. The lift-and-shift methodology requires the data, once in the cloud, to undergo an optimization phase. Optimization allows the data to run on less CPU and random-access memory, significantly lowering costs as a result. The LivIT team works closely with Amazon during the optimization phase to make decisions on how to refactor the applications and servers moved by CloudEndure.
“The migration process involves a lot of moving pieces—it comes down to understanding the services and their environments and working with other team members to find the best way of extracting the server and migrating it, and then replicating the process to other servers,” says Weaver.
This project has offered new perspectives and an advantageous top-down look at current processes and methods of operation within LivIT. Weaver and Sproba have high hopes for AWS and its ability to modernize the way LivIT operates.
Sproba adds, “Embracing the cloud means looking at the big picture and accepting that while these changes will present new challenges, overcoming these obstacles is what makes the Laboratory a leader in innovation and technology.”