Technology Leadership in Enabling Scalable Reconfigurable Computing

Accolade Technology is a United States domiciled company headquartered in Massachusetts specializing in advanced Cyber Security and Network Monitoring solutions. Accolade serves leading commercial OEM customers as well as the government sector through its patented, scalable and industry-leading reconfigurable computing technology.

Accolade’s Advanced Cybersecurity Acceleration Suite (ACAS) FPGA Offload firmware and software powers Xilinx Alveo™ Accelerators/SmartNICs, Achronix VectorPath™ Accelerators/SmartNICs, ATLAS Service Nodes and ANIC SmartNICs to deliver optimized data to Network Monitoring and Security Tools through advanced features such as 100% packet capture, flow classification, flow shunting, deduplication, header stripping, packet filtering, and much more.

Accolade Technology Milestones

Utilizing Accolade’s universal API, customers can seamlessly migrate across the entire ANIC and Xilinx Alveo product line from 1-100GE.

  • May 2021 – Block-Allow List processing – (formerly known as Black / White List processing) Block List processing is an access control mechanism that denies access to internet connections or flows based on attributes such as email addresses, IP addresses, domain names, etc. A Block List can be applied at various points in a security architecture (Firewalls, Intrusion Prevention Systems (IPS), Web Proxys, Authentication Gateways. As opposed to Block List processing, Allow List processing only allows items on the list to pass through. Real-time implementations of Block/Allow List processing such as Authentication Gateways and Firewalls are required to handle demanding, high capacity (nx100G) Network Connections while executing complex security algorithms. [more]
  • January 2021 – ATLAS-2000 Quad 100GE Service Node – The Accolade Technology ATLAS-2000 Service Node or ASN performs advanced flow and packet handling functions at 200Gbps. The ASN is designed for value-added functions that include deduplication, header stripping and packet slicing. It can be deployed in 2 different configurations:

    1) In augmentation mode it serves as an adjunct to a Packet Broker with traffic coming in from a Packet Broker and U-turned on the same port once selected functions are performed. In this mode the ASN centralizes advanced packet processing, especially with Packet Brokers based on commodity Ethernet Switches resulting in significant cost savings. In addition the ASN architecture provides increased flexibility as multiple packet processing functions can run in one ASN and new functions can be seamlessly added as required.

    2) With Inline mode the ASN sits in the traffic path and performs packet processing functions as traffic is flowing through the network. For example, in this configuration IP traffic can be filtered or scrubbed on-the-fly based on black or white lists thereby preventing malicious traffic from entering an enterprise environment. [more]
  • July & October 2020 – Powering mission critical applications for Xilinx Alveo U50 and U280 PCIe cards Accolade’s Software and Firmware power the Xilinx® Alveo™ U280 Data Center accelerator PCIe cards (SmartNICs) in advanced Cyber Security, Network Monitoring and Cloud centric applications. These features include the dynamic classification of flows, deduplication, packet slicing, block-allow list processing and much more. Built on the Xilinx 16nm UltraScale™ architecture, the Alveo U280 offers 8GB of HBM2 up to 460 GB/s1 bandwidth to provide high-performance, adaptable acceleration for other memory-bound compute intensive applications including analytics, and machine learning inference. [more]

    The Xilinx® Alveo™ U50 Data Center accelerator card powered by Accolade Software/Firmware provide optimized acceleration for workloads in applications for Cyber Security, Network Monitoring, financial computing, machine learning, computational storage, and data search and analytics. Built on Xilinx UltraScale+™ architecture and packaged in an efficient 75-watt, low-profile form factor, the U50 includes 8GB HBM2, 100GbE networking, and PCI Express 4.0 and is designed for deployment in any server. [more]
  • May 2020 – ATLAS-1100, 4 x 10GE Service Node – The Accolade ATLAS-1100 Service Node or ASN performs advanced flow management and packet handling functions at 40Gbps for Network Monitoring and Security Tools. Designed to maximize tool performance, the ASN is configurable via a user interface to perform value-added functions including deduplication, flow shunting, header stripping and packet slicing. It is deployed in two different configurations:

    1) In augmentation configuration it serves as an adjunct to a packet broker with traffic fed from a packet broker and then sent back (U-turn) after the selected function is performed. With the increased use of commodity Ethernet switches or lower cost packet brokers in networks, the ASN centralizes advanced packet processing functions such as deduplication, header stripping and packet slicing resulting in significant cost savings.

    2) With inline configuration the ASN sits in the traffic path and performs functions as traffic is flowing through the network. In this configuration security functions based on deep packet inspection (DPI) are performed “on-the-fly” to prevent unwanted traffic from entering an enterprise environment. In another option traffic enters the ASN via the 40G interface and after DPI or other processing it “fans-out” via the four separate 10G interfaces. [more]

100% Packet Capture

At the core of all ANIC adapters is the FPGA based Advanced Packet Processor (APP) which features a high performance pipe-lined packet processing engine, a high performance DDR3 memory and PCIe DMA interface.

Learn More »

Packet Filtering

At the core of all ANIC adapters is the FPGA based Advanced Packet Processor (APP) which features a high performance pipe-lined packet processing engine, a high performance DDR3 memory and PCIe DMA interface.

Learn More »

Flow Classification

The APP hash-based classifier generates a 24 bit hash key using the Layer 3/4  five-tuple (IP source address, UDP/TCP source port, IP destination address, UDP/TCP destination port, protocol id).

Learn More »

Packet Merging

In applications where a single receive port is connected to an aggregation Tap, Mirror or SPAN port of a Switch, packets are time stamped in order of arrival, processed by the APP and forwarded to the Host CPU.

Learn More »

Flow Shunting

Flow shunting allows an application to programmatically turn packet transmission on or off—for a given flow, thereby reducing the amount of data it has to analyze, freeing up CPU resources for more pressing tasks.

Learn More »

Deduplication

Most network monitoring appliances receive a significant number of duplicate packets; sometimes as high as 50% of all traffic. ANIC adapters can discard all duplicate packets in hardware before they ever reach the host application, saving wasted processing cycles.

Learn More »