Home >> Products >> 10/25/40/100G Packet Broker with PCAP Filtering

10/25/40/100G Packet Broker with PCAP Filtering

FPGA-Based PCIe Card with Advanced Network Filtering IP
  • NPP Network Packet Processing
  • HPC High Performance Computing
  • SP Signal Processing
  • NPP    
  • C&D Compute & Data Center
  • GOV Government
  • I&T Instrumentation & Test
  • FIN Financial Services
  • B&V Broadcast & Video
  • C&D GOV      

PCIe x16 card designed for integration into any system

  • QSFP-28 Ports allow for the latest IEEE 802.3 10G, 40G, and 100G pluggable modules
  • 10/25/40/100G line rates
  • PCIe DMA interface to host operating system with optional NIC interface

Advanced PCAP filter engine

  • PCAP syntax expression synthesizer
  • Deep packet inspection of all layers of the OSI model
  • Minimum term computation algorithm for maximum resource efficiency
  • Transparent processing without modifying packet data
  • Flexible network traffic flow & port mappings, many-to-many, any-to-many, and any-to-any configurations
  • Aggregation, decimation, slicing, and de-duplication
  • Line-rate policier, masking & timestamping engine
  • RMON1 statistics counters pre and post filter
  • Modular scalable design adjusts to fit any size application
  • Packet sizes up to 16Kbytes
  • Software upgradable via aftermarket licensing


The 10/25/40/100G Packet Broker with PCAP Filtering incorporates advanced packet filtering IP into a BittWare Xilinx FPGA-based PCIe card. The Packet Broker with PCAP Filtering is the industry’s first to support packet filter changes on the fly – without reconfiguring the FPGA – since filter parameters are built into the FPGA.

The PCIe x16 card can be inserted into any system to add filtering packet broker functionality, which gathers, selects, aggregates, copies, and routes network traffic to security solutions and performance analysis tools. Advanced PCAP filtering extends the capabilities of traditional packet brokers by enabling network professionals to select traffic of interest from the packet stream, relieving overburdened monitoring tools. The Packet Broker PCIe hardware can be upgraded remotely via software images, making the packet processing solution future proof. You can deploy a 10G processing solution now and upgrade to 100G later using the same hardware.

The PCIe cards can optionally be installed in a BittWare TeraBox rackmount system to create a complete packet processing appliance. The robust combination of 10/25/40 and 100G cards combined with the TeraBox allows you to build virtually any packet processing solution to fit your needs. Couple that with optional PCIe DMA or NIC interface, and it becomes one of the most feature-rich packet brokers in the industry.

Product Configurations

The packet broker has four basic product configurations:

Each of the above basic product configurations can be paired with either a PCIe DMA interface or a PCIe NIC interface, creating a total of eight product configuration options.

Product Description

As shown in the detailed block diagram, each packet presented to the card is analyzed by RMON1 counters, line rate policy, and filter-logic that determines if the frame’s data contents match a set of industry-standard PCAP ASCII expressions. The logic and PCAP ASCII expressions are not limited to the top-layers of the frame. The entire frame’s data contents are accessible including the Preamble, Start-of-Frame Delimiter, and CRC, which allows for deep packet inspection of all OSI Layers (1 through 7) at line rate. The latency of each frame is static and is proportional to the resources allocated. Frames matching the user-configured set of expressions are identified by a packet-match logical output driven synchronously with the frame on the transmit Interface.

Once packet evaluation and statistics gathering is complete, Ethernet packets are then passed into the aggregation fabric where multiple streams are combined, decimated, and/or de-dupilcated. The packets then pass out of the aggregation fabric to either another port via an additional point-to-point processing chain or to the host operating system via PCIe DMA or NIC interface.

Custom software drivers interpret the user input PCAP ASCII expressions and synthesize them into minimum terms for the most efficient use of hardware filtering resources. Resources are then divided and placed in different locations within the processing fabric to ensure optimum traffic flow and analysis. Many packet broker systems use PCAP processing expressions, but the point-to-point fabric from BittWare is the first in the industry to optimize user inputs before loading resource settings. This allows more rules to be packed into a smaller space and increases packet processing efficiency. RMON statistics counters and fabric settings are available via driver API calls.

Block Diagram

Ordering Options

Ordering Information

Contact BittWare for details.

Why BittWare?

Quality & Life Cycle Management

Technology refresh capability with 10+ year supply, quality and timely manufacturing and a trustworthy supply chain. ISO 9000 certified FPGA platform vendor with a history of proven, successful, on-time supply.  Learn more.

FPGA Board Platforms

BittWare's Virtex and Kintex UltraScale low-profile and 3/4-length PCIe board platforms support dual & quad QSFP, DDR4, configurable hardware options, data conversion I/O options, and air-cooled operation.  Learn more.

FPGA Development Tools

Designed to make developing and debugging the applications for BittWare’s boards easy and efficient, the Toolkit provides the glue between the host application and the hardware.  Learn more.

System Development & Integration

Host development tools, Board Support Packages, Baseboard Management Controller and integrated platforms including chassis, backplanes for a full solution. Learn more.

Looking for additional information?

Your Name (required)

Your Email (required)

Your Company


Your Message

Please leave this field empty.

Your Name (required)

Your Email (required)


Your Message

Please leave this field empty.