Ethernet KVS acceleration by Axonerve technology

Ethernet KVS acceleration by Axonerve technology

In this project, we are developing a high throughput KVS apps using the Ethernet interface of Xilinx FPGA cards. Designed the entire system on the Alveo FPGA card, we have achieved 150MOPS throughput and a large scale KVS system up to 64M by placing the table data to the HBM side. This application is targeting for Redis system, real-time data analysis, and security applications.

Why KVS systems by ethernet interface?

In this challenge, in order to implement high speed processing and large-scale search at the same time, the entire system is implemented by accessing data via Ethernet and placing KVS tables on the FPGA-HBM side. This system is based on Axonerve technology, that is our unique and original technology.

Design Metrix


Axonerve is just company’s inside venture unit, still looking for good partnerships even through for cloud application side. Currently, we are starting to develop to Redis acceleration solution using this KVS and Axonerve technology.

However, on the cloud application side, we do not have enough development resources and knowledge, and would like to collaborate with cloud servicers’ and software vendors who are active in these areas. We will focus on releasing platforms and reference designs that will enable ad accelerate development in this area.

Axonerve IP core technology overview

Axonerve is our original technology its an algorithmic table search engine IP core with wildcard search function. Targeting for FPGA-based applications such as Networking (Telecom Carrier, Data Center) and High-Performance Computing. #Low Latency, #High Throughput, #Hugh Entries, Wildcard Search

Kaoru Kobayashi
Marketing Manager at Axonerve technology

Ethernet KVS by Axonerve

Product presentation

Key Value Store (KVS) is a valid technique in systems such as telecommunications directories, IP transfer tables, and deduplication storage systems. In this time, we present an example of Accelerated KVS system by Alveo U50 with a lookup engine called ”Axonerve”.


On the Xilinx App Store

Share on:

Share on linkedin
Share on twitter
Share on facebook
By Jean-Michel Frouin on 11 March 2021