![](https://www.cioreviewindia.com/cioimages/PnOZClJG.jpeg)
![](https://www.cioreviewindia.com/cioimages/p5EaCTD0.jpeg)
Acceleron labs’ MDC accelerates micro-server, storage and network access. MDC comes in two form factors, MDC400 and MDC1000. Both of them are built on MDC Super Compute Fabric, which has maximum of 9.6 terabits of switching capacity and minimum 140 ns latency. MDC Super Compute Fabric delivers near-infiniband performance at costs of Ethernet. MDC has universal slots which can accommodate server/storage/GPU/DSP cartridges. MDC has facility to accommodate hybrid storage such as SATA, SAS and NVMe storage devices. Their configurable network features allow users to select combination of Ethernet, Infini-band, PCIe, etc for uplink interface.
MDCs to Enable Powerful Computing
Acceleron Labs’ current product MDC1000 has 72 Intel Xeon D (D-1529/D-1537/D-1548) processors and each processor has a maximum capacity of up to 4TB on board storage accompanied by a memory of 48 GB.
Acceleron Labs’ current product MDC1000 has 72 Intel Xeon D (D-1529/D-1537/D-1548) processors and each processor has a maximum capacity of up to 4TB on board storage accompanied by a memory of 48 GB.
All compute nodes are connected in Acceleron’s patented low latency switch fabric and MDC fabric has a unique capability to share other computing (FPGA/GPU/DSP) and storage resources across Intel server processors in the system.
MDC is a highly flexible, innovative and compact software defined data center available in the industry. Its scalability across its components is one of the USP of MDC in situations when a customer wants only 34 processors with the balance to be used as storage or a 34 processor with 38 GPU or combination of processors, storage, GPU and FPGA, Acceleron Labs’ makes it available in a customized configuration as demanded by the client. These configurations are called hyper converged hybrid compute, hybrid storage, hybrid network, and acceleration system.
In a unique proposition, the hardware being plugged in the slots are equipped to be shared across the compute node. This is achieved through switch fabric. “Due to our architecture we are able to share computing resources across multiple processors nodes (eg; 10 processors)”, L. K Prasad, MD, Acceleron Labs. MDC super compute Fabric connects FPGAs and GPUs to Compute Nodes with same BW and Latency of x8 PCIe Interface.
MDC is a highly flexible, innovative and compact software defined data center available in the industry. Its scalability across its components is one of the USP of MDC in situations when a customer wants only 34 processors with the balance to be used as storage or a 34 processor with 38 GPU or combination of processors, storage, GPU and FPGA, Acceleron Labs’ makes it available in a customized configuration as demanded by the client. These configurations are called hyper converged hybrid compute, hybrid storage, hybrid network, and acceleration system.
In a unique proposition, the hardware being plugged in the slots are equipped to be shared across the compute node. This is achieved through switch fabric. “Due to our architecture we are able to share computing resources across multiple processors nodes (eg; 10 processors)”, L. K Prasad, MD, Acceleron Labs. MDC super compute Fabric connects FPGAs and GPUs to Compute Nodes with same BW and Latency of x8 PCIe Interface.
With its team of ace electronics engineers, Acceleron Labs leveraging its MDS systems is focused on introducing high speed inter-connect , software defined hybrid computing and wireless management on data center networks.
Target following applications
• Edge Computing & Analytics
• e-Commerce Analytics
• storage Box
• Hadoop Clusters
• OpenStack
• Mini HPC clusters
• Network Function Virtualization
• AI & Machine Learning
Target following applications
• Edge Computing & Analytics
• e-Commerce Analytics
• storage Box
• Hadoop Clusters
• OpenStack
• Mini HPC clusters
• Network Function Virtualization
• AI & Machine Learning