{"id":48,"date":"2026-07-14T09:44:34","date_gmt":"2026-07-14T13:44:34","guid":{"rendered":"https:\/\/blogs.mathworks.com\/engineering\/?p=48"},"modified":"2026-07-14T09:44:34","modified_gmt":"2026-07-14T13:44:34","slug":"on-scaling-model-based-design-into-the-cloud","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/engineering\/2026\/07\/14\/on-scaling-model-based-design-into-the-cloud\/","title":{"rendered":"On scaling Model-Based Design into the Cloud"},"content":{"rendered":"
\n
For this blog, I’ve collaborated <\/span>extensively<\/span> with <\/span>Gandharv Kashinath<\/span><\/a>, who is a Principal Product Manager at MathWorks and a guru when it comes to running MATLAB in the cloud.<\/span><\/div>\n

<\/span><\/h2>\n
Have you kicked off a parameter sweep on Friday afternoon knowing full well it will not finish before Monday?<\/span> <\/span>Or watched a<\/span> <\/span>regression<\/span> <\/span>suite queue for hours because half the team is fighting over the same<\/span> <\/span>compute<\/span>?<\/span> <\/span>Now, imagine these problems across<\/span> <\/span>100<\/span>s of<\/span> <\/span>engineers<\/span>.<\/span> <\/span>This is<\/span> <\/span>exactly<\/span> <\/span>the issue modern OEMs are facing right now. <\/span> <\/span><\/div>\n
What\u2019s<\/span> <\/span>accentuating these issues is the shift to s<\/span>oftware-defined vehicles (SDVs)<\/span>.<\/span> <\/span>Functionality<\/span>,<\/span> <\/span>from powertrain control to ADAS and connectivity<\/span>,<\/span> <\/span>is increasingly delivered through software, with continuous updates across the vehicle lifecycle.<\/span> <\/span>Every update means<\/span> <\/span>more<\/span> <\/span>simulation<\/span>s<\/span>; regression pass<\/span>es; and<\/span> <\/span>sweep<\/span>s<\/span>.<\/span> <\/span><\/div>\n
This shift is changing how engineering teams<\/span> <\/span>operate<\/span>.<\/span> <\/span>The same engineer might need to run simulations, generate production code, and pull in fleet data for designing and validating algorithms, often stitching together tools, teams and sometimes time zones to do so.<\/span> <\/span>Model-Based Design<\/span> <\/span>is central<\/span> <\/span>to this process. MATLAB and Simulink<\/span> <\/span>allow<\/span> <\/span>teams<\/span> <\/span>to achieve<\/span> <\/span>all<\/span> <\/span>the<\/span> <\/span>above<\/span> <\/span>with<\/span> <\/span>a workflow<\/span> <\/span>that provides a digital thread from requirements through verification and validation.<\/span> <\/span><\/div>\n
What\u2019s<\/span> <\/span>changing now is the scale, how much complexity,<\/span> <\/span>validation<\/span> <\/span>and data these workflows<\/span> <\/span>must<\/span> <\/span>carry, especially in SDV<\/span> <\/span>programs<\/span>.<\/span> <\/span>In this post, we<\/span> <\/span>are discussing<\/span> <\/span>how engineering teams are extending these workflows. Specifically, we show how MATLAB and Simulink, combined with<\/span> <\/span>cloud<\/span> <\/span>and data platforms<\/span>,<\/span> <\/span>such as AWS and Databricks,<\/span> <\/span>are helping distributed teams run these workflows consistently and at scale.<\/span> <\/span><\/div>\n
\"\"<\/img> <\/span><\/div>\n
Figure 1.<\/span> <\/span>Software-defined vehicle development with Model-Based Design<\/span> <\/span>in the Cloud. IT vs. Engineering priorities.<\/span> <\/span><\/div>\n
<\/span><\/div>\n

Why scaling is hard<\/span> <\/span><\/h2>\n
Scaling is hard because<\/span> <\/span>engineering teams want to<\/span> <\/span>run bigger<\/span> <\/span>simulations<\/span> <\/span>and<\/span> <\/span>pull in<\/span> <\/span>an ever-<\/span>increasing<\/span> <\/span>amount of<\/span> <\/span>telemetry<\/span> <\/span>data, whilst IT teams need to manage<\/span> <\/span>access, enforce<\/span> <\/span>security and governance and provision<\/span> <\/span>compute<\/span> <\/span>resources.<\/span> <\/span>What this can mean in practice<\/span> <\/span>is engineer<\/span>s<\/span> <\/span>waiting on a ticket<\/span> <\/span>for access to compute<\/span> <\/span>while a regression suite queues, or a supplier team waiting on access and setup before they can even start comparing results against the OEM<\/span>.<\/span> <\/span>We\u2019ve<\/span> <\/span>seen these play out across automotive OEMs.<\/span><\/div>\n
<\/span><\/div>\n
Three areas commonly limit scalability.<\/span> <\/span><\/div>\n
    \n
  • Simulation workloads exceed local compute capacity<\/span>: <\/span>R<\/span>egression<\/span> <\/span>testing, parameter sweeps, and system-level simulations<\/span> <\/span>routinely<\/span> <\/span>outgrow<\/span> <\/span>a<\/span> <\/span>workstation<\/span> <\/span>and running them locally means queuing for IT-managed<\/span> <\/span>compute<\/span>.<\/span> <\/span><\/li>\n
  • Data is difficult to connect and reuse:<\/span> Simulation outputs, test results, and telemetry data are often<\/span> <\/span>scattered<\/span> <\/span>across tools and<\/span> <\/span>teams<\/span> <\/span>making it harder to keep requirements,<\/span> <\/span>models<\/span> <\/span>and tests traceable<\/span>.<\/span> <\/span><\/li>\n
  • Distributed teams <\/span>require<\/span> <\/span>consistent environments<\/span>:<\/span> Suppliers need to have access to<\/span> <\/span>the same<\/span> <\/span>tooling<\/span> <\/span>and compute<\/span> <\/span>setup<\/span> <\/span>as the OEM.<\/span> <\/span>Getting everyone aligned<\/span> <\/span>adds overheads that can slow development. <\/span><\/li>\n<\/ul>\n
    <\/span><\/div>\n

    Scaling Model-Based Design<\/span><\/h2>\n
    In recent engagements with OEMs, we\u2019ve supported them in extending existing workflows with cloud and data platforms without changing how they work day to day. We set them up with MathWorks Cloud Center so that they can create, access and manage public cloud resources themselves, allowing engineering to keep moving forwards while IT retained control over what was provisioned. <\/span><\/div>\n
    Getting started is not hard. All you need is an AWS account & then you add your credentials in Cloud Center. From there, you can spin up an AWS instance with MATLAB already installed. <\/span><\/div>\n
    <\/span><\/div>\n
    For more detail, my colleague Tianyi <\/span>Zhu<\/span> <\/span>explains the full setup in this short video.