Packet loss detection in real-time communications
A method includes determining a packet loss rate in a real-time communication event based on a size of a packet block in a received sequence comprising multiple packets, determining a number of packet loss clusters within the packet block based on the packet loss rate, identifying one of a random loss pattern or a burst loss pattern based on the number of packet loss clusters or based on whether packet loss pattern matches a geometric distribution, including an error correction bit in at least one of the packets in the packet block when a random loss pattern is identified, and reducing a packet transmission rate when a burst loss pattern is identified. A system including a memory storing instructions which, when executed by a processor, cause the system to perform the above method, is also provided.
1 . A computer-implemented method, comprising:
determining a packet loss rate associated with a real-time communication event based on a size of a first packet block;
determining a number of packet loss clusters within the first packet block based on the packet loss rate;
identifying, based on the number of packet loss clusters, a first loss pattern indicative of a random packet loss;
in response to identifying the first loss pattern, including an error correction bit in a packet in the first packet block;
identifying a second loss pattern in a second packet block, wherein the second loss pattern is indicative of network congestion; and
in response to identifying the second loss pattern, reducing a packet transmission rate in the second packet block;
wherein identifying the first loss pattern or the second loss pattern is based on an average packet distance between packet loss clusters in the first packet block.
2 . The computer-implemented method of claim 1 , wherein determining the packet loss rate further comprises determining that a packet is a retransmitted packet.
3 . The computer-implemented method of claim 1 , further comprising selecting the size of the first packet block or second block and forming a test window that includes at least two packet blocks.
4 . The computer-implemented method of claim 1 , wherein determining the packet loss rate further comprises determining a number of packets lost within the first packet block.
5 . The computer-implemented method of claim 1 , wherein a packet loss cluster comprises a plurality of lost packets that are separated by fewer than a pre-determined number of packets.
6 . The computer-implemented method of claim 1 , wherein determining the number of packet loss clusters is based on a total packet loss in the first packet block or second packet block.
7 . The computer-implemented method of claim 1 , wherein identifying the first loss pattern further comprises determining that a packet loss distribution follows a geometric distribution.
8 . The computer-implemented method of claim 1 , wherein identifying the second loss pattern further comprises verifying that the packet loss rate is higher than a threshold value.
9 . The computer-implemented method of claim 1 , wherein identifying the first loss pattern or the second loss pattern is based on an average packet distance between missing packets within a packet loss cluster.
10 . The computer-implemented method of claim 1 , further comprising reducing the size of the first packet block or second packet block for a selected number of blocks.
11 . The computer-implemented method of claim 1 , further comprising increasing the size of the first packet block or second packet block for a selected number of blocks.
12 . A system, comprising:
a memory storing multiple instructions; and
one or more processors configured to execute the instructions to cause the system to:
determine a packet loss rate associated with a real-time communication event based on a size of a first packet block;
determine a number of packet loss clusters within the first packet block based on the packet loss rate, wherein a packet loss cluster comprises a plurality of lost packets that are separated by fewer than a pre-determined number of packets;
identify, based on the number of packet loss clusters, a first loss pattern indicative of a random packet loss;
in response to identifying the first loss pattern, include an error correction bit in a packet in the first packet block;
identify a second loss pattern in a second packet block, wherein the second loss pattern is indicative of network congestion; and
in response to identifying the second loss pattern, reduce a packet transmission rate in the second packet block.
13 . The system of claim 12 , wherein determining the packet loss rate further comprises determining that a packet is a retransmitted packet rate.
14 . The system of claim 12 , wherein the one or more processors further execute instructions to select the size of the first packet block or the second packet block and forming a test window that includes at least two packet blocks.
15 . The system of claim 12 , wherein determining the packet loss rate further comprises determining a number of packets lost within the first packet block.
16 . The system of claim 12 , wherein determining the number of packet loss clusters within the first packet block or the second packet block based on a total packet loss in the first packet block or the second packet block.
17 . The system of claim 12 , wherein identifying the first loss pattern further comprises that a packet loss distribution follows a geometric distribution.
18 . A computer-implemented method, comprising:
determining a packet loss rate associated with a real-time communication event based on a size of a first packet block;
determining a number of packet loss clusters within the first packet block based on the packet loss rate, wherein the packet loss cluster comprises a plurality of lost packets that are separated by fewer than a pre-determined number of packets;
identifying, based on the number of packet loss clusters, a first loss pattern indicative of a random packet loss;
in response to identifying the first loss pattern, including an error correction bit in a packet in the first packet block;
identifying a second loss pattern in a second packet block, wherein the second loss pattern is indicative of network congestion; and
in response to identifying the second loss pattern, reducing a packet transmission rate in the second packet block.