Network Performance Analytics Startup Relies on ATLAS-1000
Proprietary Hash Algorithm Calculation in the FPGA
Resulted in 70% Savings in CPU Cycles
Learn how a network performance analysis startup, laser focused in developing breakthrough analytics solutions, solved key challenges to deliver a distinct market leading product. The platform had to meet specific requirements to be deployed in multiple telecom customer networks as a network sensor or probe.
Specifically for this startup, Accolade implemented a proprietary hash algorithm calculation in the FPGA, resulting in an eye-popping 70% savings in CPU cycles. This particular custom feature resulted in significant performance improvement and was one of the driving forces behind selecting Accolade for this time sensitive project.
Key Challenges:
- Highly CPU intensive hash algorithm
- Ultra precise timing
- Platform requiring small footprint
- Lossless packet capture across all ports
- 10G and 40G network interfaces with optical tap
ATLAS-1000 Solution:
- Hash processing algorithm in FPGA = 70% savings in CPU cycles
- Hardware timestamping with OCXO assist
- Two ATLAS-1000 platforms fit in 19 inch rack
- 100% packet capture
- 10G and 40G network interfaces with integrated optical tap