Optimizing transfer of fragmented packetized data

A non-transitory computer-readable storage medium having embodied thereon a program executable by a processor to perform a method for optimizing the transfer of packetized data between electronic devices in a computer network, the method comprising:

• receiving data packets of a set comprising a plurality of fragmented internet protocol (IP) packets sent from a source computer to a destination computer for processing at the destination computer, the received set of fragmented IP packets comprising a number of bytes;
• identifying that the set of fragmented IP packets includes redundant information;
• assembling a first number of the set of fragmented IP packets into a coalesced IP packet, wherein assembling the set of fragmented IP packets comprises removing at least a portion of the redundant information to reduce the number of bytes, and adjusting a header of the coalesced IP packet to include a length of the coalesced IP packet;
• compressing the coalesced IP packet;
• identifying a size of the compressed coalesced IP packet, the size of the compressed coalesced IP packet corresponding to a first average size;
• identifying that the size of the compressed coalesced IP packet is less than a maximum transmission unit size;
• transmitting the compressed coalesced IP packet to the destination computer;
• identifying a second number of fragmented IP packets to coalesce into a second coalesced IP packet based on comparing the first average size to the maximum transmission unit size and increasing the number of the set of fragmented IP packets to include in the second coalesced IP packet to a second number of packets according to a calculation;
• assembling the second number of the set of fragmented IP packets in to a second coalesced IP packet;
• compressing the second coalesced IP packet;
• and
• transmitting the second compressed coalesced IP packet to the destination computer.

A system for optimizing the transfer of packetized data between electronic devices in a computer network, the system comprising:

• one or more computer network interfaces that:
• receive data packets of a set comprising a plurality of fragmented internet protocol (IP) packets sent from a source computer to a destination computer for processing at the destination computer;
• a memory;
• and
• a processor that executes instructions stored in the memory, wherein execution of the instructions:
• identifies that the set of fragmented IP packets includes redundant information,
• assembles a first number of the set of fragmented IP packets into a coalesced IP packet, wherein assembling the set of fragmented IP packets comprises removing at least a portion of the redundant information to reduce the number of bytes, and adjusting a header of the coalesced IP packet to include a length of the coalesced IP packet, and
• compresses the coalesced IP packet,
• identifies a size of the compressed coalesced IP packet, the size of the compressed coalesced IP packet corresponding to a first average size,
• identifies that the size of the compressed coalesced IP packet is less than a maximum transmission unit size, wherein one of the one or more computer network interfaces transmits the compressed coalesced IP packet to the destination computer,
• identifies a second number of fragmented IP packets to coalesce into a second coalesced IP packet based on comparing the first average size to the maximum transmission unit size and increasing the number of the set of fragmented IP packets to include in the second coalesced IP packet to a second number of packets according to a calculation, and
• assembles the second number of the set of fragmented IP packets into a second coalesced IP packet, and
• compresses the second coalesced IP packet, wherein the second compressed coalesced IP packet is transmitted to the destination computer.