Partners and Technologies

At Grid Dynamics Consulting Services, we solve customer problems with our demonstrated expertise in all the relevant technologies and by partnering with world-class organizations:

Data Grids

Data Grids allow large clusters of systems to support an application running with a large memory footprint. They provide superior levels of scalability, flexibility, and reliability, and are a great complement to Compute Grid technologies. We have the deep expertise to help customers evaluate, select, implement, integrate, and maintain data grid middleware. We have had several contracts deploying the Space Based Architecture from GigaSpaces, a particularly powerful solution for clusters of 100s or 1000s of instances of Java virtual machines working with shared online cached spaces. We have evaluated and deployed systems based on other leading technologies, including the Gemstone Enterprise Data Fabric and Tangosol Coherence.

Compute Grids

Compute Grids support applications that require massive computational resources or high throughput, low latency information processing. We can enable our customers to evaluate, select, implement, integrate, and maintain compute grid technology, including Platform Computing, DataSynapse, and Sun Microsystems. Where we excel in this space is our ability and experience to seamlessly integrate both compute grids and data grids to implement the fastest, most scalable applications possible.

Java

Although we are skilled in C, C++, C#, and .Net, Java is used throughout many of our projects. Java offers the advantages of portability and ease of development time due to its inherent object-oriented model and use of standardized libraries for most mainstream computing tasks. Encapsulation, polymorphism, and inheritance allow for rapid and reliable development of applications. We specialize in high performance applications, and we are familiar with real-time constraints on the design of Java, including tuning of the heap (garbage collection).

A very natural extension of Java for high performance Tuple-space architectures is JavaSpaces. JavaSpaces is used to achieve scalability through parallel processing and provides for reliable storage of objects while reducing the complexity of traditional distributed systems.

Linux/Solaris/Windows

Although the Java platform means portability for application code, it is still necessary to maintain experience in all of the key platform technologies. We have deep knowledge of the kernels of these operating systems, as well as their support for framework toolkits such as C, C++, C#, and .Net.

Networking/Infrastructure

High throughput solutions depend on chasing latencies right out of the system. Systems must be designed for low latency throughout the networking fabric (Infiniband, 10Gb Ethernet, and proprietary interconnects from HPC) as well as the compute nodes. It is essential to know how to optimize these solutions when adding online storage protocols, including storage NAS and SAN protocols, as well as understanding the detail behind host and target adapters.