Back to results
US 10,715,618 B2Granted
Compressibility estimation for lossless data compression
AssigneeHughes Network Systems, LLC
InventorUday R. Bhaskar
Life of this patent
Estimating the term…
Abstract
A network node includes a processor programmed to parse at least a portion of an input block having a plurality of segments, determine whether at least one of the plurality of segments matches a segment stored in a history buffer, and predict a compressibility of the input block based at least in part on whether at least one of the plurality of segments matches a segment stored in the history buffer.
Claims15
A network node comprising:
- a memory;
- and
- a processor programmed to execute instructions stored in the memory, the instructions including:
- parse at least a portion of a data packet, wherein the packet comprises at least a first segment and a second segment;
- determine a first token representing the first segment;
- using the first segment, calculate a cumulative bit cost;
- based on the cumulative bit cost and at least one of:
- a current number of parsed bytes of the packet, a total packet length, or a current number of tokens associated with the packet, predict either to transmit the packet in a compressed state or to transmit the packet in an uncompressed state;
- compress the packet into the compressed state if the cumulative bit cost is less than a bit cost threshold during parsing of the packet;
- determine a second token representing the second segment;
- using the second segment, calculate an updated cumulative bit cost;
- based on the updated cumulative bit cost and at least one of:
- an updated current number of parsed bytes of the packet, the total packet length, or an updated current number of tokens associated with the packet, re-predict to either transmit the packet in the compressed state or to transmit the packet in the uncompressed state;
- and
- when the prediction is to transmit the packet in the compressed state, then perform arithmetic encoding on the first and second tokens.
12 more elements
A method, comprising:
- parsing at least a portion of a data packet, wherein the packet comprises at least a first segment and a second segment;
- determining a first token representing the first segment;
- using the first segment, calculating a cumulative bit cost;
- based on the cumulative bit cost and at least one of:
- a current number of parsed bytes of the packet, a total packet length, or a current number of tokens associated with the packet, predicting either to transmit the packet in a compressed state or to transmit the packet in an uncompressed state wherein predicting a compressibility of the packet is based at least in part on a comparison of the current number of parsed bytes of the packet relative to the cumulative bit cost;
- and
- after parsing the packet, compressing the packet to the compressed state by performing arithmetic encoding on a plurality of tokens generated during parsing of the packet if the cumulative bit cost determined while parsing the packet is less than a bit cost threshold and the plurality of tokens generated while parsing the packet is less than a token threshold, wherein the plurality of tokens comprise the first token and a second token associated with the second segment,
- wherein the bit cost threshold is based at least in part on the current number of parsed bytes and wherein the token threshold is based at least in part on the plurality of tokens generated as a result of parsing the packet.
5 more elements
Dependencies are inferred from the claim text.