System Analysis and Design
CLUSTERTECH offers system analysis and design services by focusing on the economic value, suitability, manageability and extensibility of the solution. With our expertise in supercomputing technologies, we offer services in designing scalable, robust and cost-effective HPC, grid computing and high-throughput I/O systems.
Examples of our work in this area include:
- High-level/functional analysis of current systems to identify areas where applications of parallel and distributed computing or high-throughput I/O systems may yield a performance boost.
- Performance benchmarking at both application and system level for the purpose of capacity planning.
- Analysis of business and operation requirements, assessment of technical feasibility as well as evaluating the deployment effort of deploying supercomputing technologies.
- Design of customized HPC and grid computing infrastructures for applications with high computational demands.
- Design of high-throughput storage systems for I/O-intensive and large-scale distributed applications.
- Design of high availability systems, including failover-capable servers, fault-tolerant I/O subsystems and multi-path interconnect, for mission-critical applications.
- Integration planning for HPC, grid computing and high-throughput I/O systems.
- Prototyping for proof-of-concept study.
System Support and Optimization
CLUSTERTECH has teams of highly skilled engineers, offering a wide range of services in supporting and optimizing HPC clusters, grid infrastructures and high-throughput I/O systems. Services include:
- Software maintenance and upgrade
Technologies in HPC and grid computing change rapidly. We make the benefits of the latest technology available to our clients by performing regular upgrade and maintenance work on their software infrastructure.
- System optimization
HPC clusters or a grid can be tuned in an application specific manner. This type of tuning can dramatically improve computational efficiency and throughput. Clustertech offers system optimization services with a focus on resource management systems, interconnect and I/O systems. This will help you to fully utilize your supercomputing investment.
- Open-source software support
The HPC community has used open-source software components extensively. While this can significantly decrease the development effort, the compilation, installation and customization, open-source components are often problematic due to the fact that every HPC system make use of different components (Linux kernel, drivers, compilers, MPI libraries, etc.). Our engineers have extensive experience in troubleshooting problems associated with incorporating open-source components in applications. |
