The accelerated driver has been optimized for extreme packet capture speed, by redesigning the packet path from the network adapter to the packet capture applications. Unnecessary memory allocations/free have been removed as well as costly data exchange over the PCI bus. Furthermore a direct mapping of the network adapter offers to monitoring applications the ability to grab packets out of the network adapter without any kernel or main CPU intervention. This means that:

• The ASIC chip on the ethernet card does all, as it is programmed to transport captured packets directly to the enhanced libpcap. Both the kernel and the main host CPU (e.g. Intel Pentium™ or AMD Athlon™) are completely bypassed during the packet capture process, resulting in extreme packet capture performance.
• As the kernel is bypassed, there is almost no latency for packets when moving from the network adapter to the monitoring application, resulting in more precise packet timestamps.
• The packet capture speed is limited only by the ability of the packet capture application to "consume" packets.
• Captured packets are not copied at all and they are always captured at full size and not just the packet header as libpcap does. In fact as the kernel and the CPU are completely bypassed, capturing full packets is not more costly as capturing just the packet header as the work is carried on by the ASIC chip on the ethernet card.

The enhanced driver is completely transparent application. This means that nCap-aware applications can fully take advantage of the driver, whereas unaware applications continue to operate as before. The driver allows to mix on the same adapter both legacy and nCap-aware applications at the same time.

At present, the enhanced network device driver is available for both Intel 1 Gbit network adapters over PCI 32/64 and PCI-X.