Kratos Defense & Security Solutions has been granted a patent for systems, media, and methods that determine a layer 2 destination address for network acceleration applications. The technology involves a software application executed on a computing device that controls a NIC interface to provide accelerated networking. The method includes creating a phantom datagram, populating protocol headers, submitting it to the kernel, receiving it back with the next-hop MAC address populated, and storing the address without transmitting the phantom datagram. This technology can account for firewall rules, static ARP entries, and other network operation rules. GlobalData’s report on Kratos Defense & Security Solutions gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on Kratos Defense & Security Solutions, drone jamming technology was a key innovation area identified from patents. Kratos Defense & Security Solutions's grant share as of September 2023 was 78%. Grant share is based on the ratio of number of grants to total number of patents.
Accelerating packet transmission by determining next-hop mac address
A recently granted patent (Publication Number: US11750516B2) describes a method and system for accelerating the transmission of packets in a network. The method involves programmatically determining the next-hop MAC address for packets to be transmitted by an application residing at the application layer in a system with a kernel and protocol stack.
The method begins by creating a phantom datagram (UDP) with protocol headers, leaving the next-hop MAC address header unpopulated. This phantom UDP is then submitted to the kernel for traversal through the protocol stack. Upon receiving the phantom UDP from the kernel, the next-hop MAC address header is populated by the kernel. The phantom UDP is recognized based on predetermined characteristics, and the next-hop MAC address header is stored. However, the phantom UDP is not transmitted.
In addition to the above steps, the method includes creating a next UDP and retrieving the stored next-hop MAC address. The header of the next UDP is then populated with the retrieved next-hop MAC address. The protocol stack of the kernel determines the next-hop MAC address header based on a predetermined set of rules that account for routing rules, network-rules, security measures, and firewall measures.
The application also has control over a network interface controller (NIC). The method includes instructing the NIC to drop the phantom UDP. Furthermore, the method allows for bypassing the kernel and providing the next UDP directly to the NIC with the stored next-hop MAC address header populated.
The method involves using a bind call and a connect call with the kernel to create a socket. The source IP and source port from the phantom UDP, which were populated by the kernel, are retrieved and stored. The header of the next UDP is then populated with the stored source IP and source port.
The granted patent also describes a system for accelerated packet transmission. The system includes at least one processor and a memory storing instructions for executing an application residing at the application layer. The application interacts with the operating system's kernel and protocol stack to create and process phantom packets, determine the layer 2 destination address, and populate the headers of subsequent packets.
Overall, this patent presents a method and system for accelerating packet transmission in a network by programmatically determining the next-hop MAC address and layer 2 destination address. The approach involves utilizing the kernel's protocol stack and allowing the application to control the network interface controller for improved efficiency.