Addressing Edge Computing Challenges with Kalray's Solutions
Edge Computing is a distributed computing technology domain that brings applications closer to data sources generated by people, machines, or objects.
According to Deloitte Digital, in the last 2 years, humans and machines have created 9 times more data than since the beginning of humanity and data volumes will continue to grow as 5G networks increase the number of connected mobile devices.
The needs for Edge Computing are very important: in industry, surveillance, smart cities, hospitals, telecommunications / 5G for drones, robots, gateways, etc… and in the data centers themselves, for example in storage servers.
Our Solutions for Edge Computing
Vision systems for industrial applications implies low cost, acceptable accuracy, high robustness and high reliability.
As most of the industries, deep learning vision software is getting a natural trend to automate inspection and make it more efficient. AI provides new opportunity for more complex inspection than traditional rule-based machine vision.
The telecom infrastructure is facing a major (r)evolution with a brand new network architecture on the Edge: Open RAN, « Software-defined » Network et Stockage as well as need for Disaggregation and open interface. Acceleration cards will play a key role: for Radio Acceleration, Network Acceleration, Application Acceleration.
A New Generation of Processors to Address Edge Computing Challenges
A new types of processor is required to capture and analysis on the fly a massive amount of information, close to where the data is generated. This is what Kalray have been developing the past 10 years with its MPPA® DPU (Data Processing Unit) processor family.
MPPA® DPU Key Benefits
MPPA® DPU (Data Processing Unit) processors capture & analyze massive data flows on the fly.
MPPA® DPU processors interact in real time with the outside world.
Kalray’s MPPA® DPU processors offer optimized & customizable power system integration
Kalray’s MPPA® DPU processors run in parallel complex/demanding AI algorithms, control tasks and other processes.