MongoDB is the leading modern, multi-platform document store, used by cutting-edge startups, various large companies and government agencies both as a fully managed cloud service and for deployment on self-managed infrastructure. Yet, there are tasks for which MongoDB as any other databases is less than optimal, for instance when processing becomes more complex or the data is less structured.
As data is exploding exponentially only CPU based systems are no longer capable of providing real-time insights to businesses in the sense of cost-efficiency.
At Grovf we offer a MongoDB acceleration platform based on Grovf MonetX – an FPGA based smart memory extension for near memory data processing. Due to this solution, MongoDB achieves 3.5X acceleration for all stages of data aggregation.
- 10GB/s, 2TB RAM memory on a single PCIe slot
- Network accessible memory
- Near memory data computing
- RAM Memory extension without increasing number of servers’ sockets
- Linux compatible
MonetX implemented on top of Xilinx’s Alveo U50 acceleration card and once plugged into the server's PCIe bus acts as a standard RAM memory for the Linux operating system with near memory compute API capability. The latest provides a simple API to host layers for easy utilization of the functions. The operating system, recognizing MonetX as a standard memory extension, also provides high-performance computing cores API for the host layer.
Check out the further technical details via datasheet.
As a result, MongoDB performance has been boosted by 3.5X only using the MonetX acceleration platform as a high bandwidth memory extension for standard server architecture without using any built-in high-performance computing cores in the FPGA. This leads to zero code change in the application(MongoDB) side and provides 3.5X acceleration. Much more acceleration for MongoDB and any other application can be achieved using built-in accelerated computing cores in the FPGA residing near memory.
Learn more about MonetX IP.