Scala Computing provides a unique solution to run discrete event-driven ns-3 network simulations for data center operators and networking equipment manufacturers with simplified on-demand access to scalable, secure compute clusters on the cloud.
Data center operators
Run what-if simulations with new networking stacks, workloads and network vendor hardware
Network Semiconductor Vendors
Test new silicon development with hyperscaler workloads before committing to tapeout
Network System Vendors
Test network silicon vendor solutions with various design options testing with end-customer workloads
Our scalable cloud-based solution allows users to scale simulations by a factor of 1000 while improving simulation performance by a factor of 10. Traffic and performance on varying workloads can be evaluated to tune your networking queuing models. It makes it easy to observe and log results that are difficult to capture in physical networks so functions and architectures can be validated or negated early and applied to next generation devices or device spins.
Pre-Processing Engine
UI-tailored for discrete event-driven network simulations (modify networking equipment, topologies, traffic and stack parameters) with automatic topology validation / debugging.
Performance
Simulations optimized for massive parallelization, enabling data center scale simulation with up to hundreds of thousands of endpoints (runtimes reduced from weeks to days and days to hours).
Data Analytics
Data analytics engine processes 100s of gigabytes of data in real-time, providing insights on latency and goodput while offering embedded performance benchmarking.
The Scala Compute Platform (SCP) provides a secure infrastructure for the orchestration, provisioning, and deployment of public cloud resources. As the network becomes more complex, network silicon testing and simulation is a critical tool to validate network design and performance. SCP for ns-3 allows for Data Center Scale simulation on the Cloud from 10s to 100s of thousands of endpoints. Additionally, the intuitive workbench provides an easy to use interface for creating your ns-3 network simulations. You can build your infrastructure on validated components from the Multi-Party Data Center Simulator Platform (MDSP) Consortium.
Understand
Highest accuracy insights about existing system networks
Predict
Predict performance of networks or components not yet deployed or available
Debug
Debug and optimize massively parallel deep learning models at scale
Optimize
Optimize network configurations and data center deployments
Improve
Improve queue management techniques and behavior of network forwarding strategies
Development Environment
For developers and contributors to build and test their own ns-3 models
