Project snapshot
Marvel Fusion is a sustainable technology trailblazer developing laser-driven fusion as a solution for the global energy transition to zero carbon emissions. Taking advantage of recent breakthroughs in the fields of laser technology and nanotechnology, Marvel’s team of scientists is working towards the commercialization of fusion energy, by utilizing ultrafast lasers and nanostructured fuel.
By applying the proprietary Intellias Systems Migration service and DevOps practices, Intellias helped Marvel Fusion to migrate to AWS cloud infrastructure resulting in improved performance, advanced high-load scientific research, and reduced time to results.
Fusion is the addition to renewable energies like solar, wind, or hydroelectric power and is capable of becoming a safe, flexible, and sustainable solution for both consumers and emission-heavy industries. Marvel Fusion’s unique laser-driven fusion technology has been validated by internationally renowned scientists.
Business challenge
Being at the forefront of ultimate clean energy development, Marvel Fusion’s team is using computer modeling to optimize and test fusion technology powered by ultra-advanced lasers.
The client’s existing setup in another cloud, combined with the local data center, did not allow them to achieve the desired level of compute power. Additionally, it was too expensive, as they had to pay steady operational expenses regardless of whether they used the servers or not.
In order to optimize performance, advance their high-load scientific research, and reduce their time to results, they were in critical need of high-performance computing opportunities that AWS cloud infrastructure could offer.
Applying DevOps practices as part of the AWS migration helped to design and migrate existing solutions smoothly. We relied on the AWS Well-Architected Framework to implement a reliable, secure, and performance-efficient solution.
Technology solution
To help design and run large-scale simulations as well as store and process the data obtained from them, we applied the proprietary Intellias Systems Migration service, which enabled Marvel Fusion to migrate to AWS infrastructure.
In addition to migration, the Intellias team also launched dedicated AWS VMs for individual workspaces to help the client’s internal team process data generated by the cloud computing infrastructure with the utmost efficiency, on top of cost and time benefits.
In close collaboration with the client, Intellias experts provided meticulous technology consulting, namely: analyzing and defining key project goals and components along with a migration roadmap, followed by a system redesign with the Well-Architected Framework. The solution was implemented by high-profile DevOps engineers in accordance with the AWS Well-Architected pillars.
Working together with software engineering teams, DevOps specialists assisted in defining infrastructure and deployment requirements, as well as performing system troubleshooting and problem resolution across various application domains and platforms.
The final architecture utilized:
- AWS Parallel Cluster for running HPC simulation workloads.
- High-performance FSx Lustre file systems for simulation scratch data storage.
- EFS for mid-term data storage with lifecycle policies to transition data to S3.
- On-demand “VM for sim evaluation” for researchers to process data using EC2 instances.
- Comprehensive CI/CD pipelines for infrastructure and application deployment.
After the Intellias DevOps experts completed the migration, they provided extensive knowledge sharing to support the Marvel Fusion team in further handling the solution with the desired level of expertise.
Applying AWS DevOps practices allowed for a hands-on approach to the overall delivery and infrastructure automation, as well as end-to-end support of continuous integration and deployment.
Business impact
By applying a tailored AWS Cloud Migration Acceleration Program, Intellias contributed to further development of Marvel Fusion’s revolutionary solution for carbon-free energy.
Our input enabled the client to leverage the benefits of AWS cloud Infrastructure in order to run simulations leading to more accurate metrics, deeper data analytics, a streamlined overall performance, and boosted the team’s productivity, thus optimizing costs and radically reducing time to market for the client’s solution.
With DevOps as an integral part of the AWS cloud migration, we set the optimal deployment cadence for managing development capabilities in the context of overall services delivery, ensuring the infrastructure remains secure, scalable, and highly available across the platforms.
The collaboration between the DevOps team and AWS solutions architects was key to the architecture and implementation of a high-performance, cost-optimized solution. Additionally, comprehensive knowledge transfer enabled the client’s researchers to effectively utilize the new platform.
By leveraging DevOps automation and following AWS best practices, the solution provided:
- Up to 15% faster execution of simulations compared to on-premises solutions
- Ability to rapidly scale to over 10,000 cores on-demand for simulations
- Costs below $0.02 per core-hour by leveraging spot instances
- Secure, auditable access and cost visibility through tagging and AWS Organizations
- Streamlined researcher workflow through self-service provisioning
Employing AWS DevOps practices accelerated performance, allowed for continuous solution improvement, and optimized expenses.