Our integration services
The whole can be more than the sum of it’s parts.
Our solutions deliver fully integrated systems incorporating hundreds of individual hardware and software components.
Compute clusters built to meet your requirements
We deploy customized systems on commodity hardware.
In 2009, we began providing an HPC cluster integration service to allow customers to deploy customized systems built with commodity hardware and software components. Following a number of customer successes, we’re taking on more resources and developing new partner relationships to deliver the best in commercial and open-source solutions.
Harness the power while keeping a lid on your costs
Linux HPC clusters provide an excellent way to harness the computing power of commodity X86-compatible rack-based server and blade servers. A cluster is often composed of a number of compute nodes with a set of dedicated machines for management, storage and interactive services. A high-performance interconnect is usually included providing high-bandwidth, low-latency communication between the cluster nodes.
You specify it. We build it.
High performance computing clusters can run a wide variety of software, packaged into a collection called the HPC cluster software stack. Your choice of stack will influence the applications your cluster can provide, the hardware on which it is built and the overall cost of the solution. Common components of a stack include:
- A 64-bit capable Linux based operating system
- A message-passing interface (MPI) for high speed node-to-node communication
- Scheduler software to control how jobs are run across the cluster
- Management software for deploying and rebuilding compute nodes
- A software development environment including compilers, debuggers, profilers
- Shared storage service for storing jobs and results
A number of predefined commercial HPC cluster stacks are available to help install a known-good set of software components suitable for creating a basic, general purpose HPC cluster. We can also provide an open-source HPC solution, combining the best freely available products with enterprise class support services. However, to realize its full potential the software stack must be carefully configured specifically for the target hardware - only when each individual component is properly tuned will maximum performance be achieved.
We work with you to design and deliver the system you need
We specialize in advising on software stack components to allow customers to create a tailor-made solution from the best commercial vendors and open-source projects. We can install, configure and demonstrate the cluster before running acceptance tests to confirm proper operation of the hardware and software stack. All our processes are fully customizable allowing you maximum control over the system delivery.
Connecting the dots
High performance computing often leverages specialised technologies long before they are accepted into mainstream IT. Deploying such solutions requires knowledgable engineers to achieve the best performance and advanced software to deliver optimum stability and usability for your customers.
We maintain a high level of expertise in key technologies including:
- Scientific software packages and applications
- System availability and performance management solutions
- Web services software and remote access systems
- Batch scheduler configuration and optimisation
- High-availability hardware and software solutions
- Distributed system performance and availability management
- Linux operating systems and package management
- InfiniBand switch and cabling technologies
- High bandwidth and low latency 40- and 10-gigabit ethernet
- High performance parallel filesystem technologies
- Special purpose hardware devices
- Commodity X86 rack and blade server systems
- Storage, backup and archive systems
- High density power distribution systems
- Liquid cooling solutions and data-centre airflow analysis
- Complex project management