Ian McDonald | 98069ff | 2005-11-10 13:04:33 -0800 | [diff] [blame] | 1 | DCCP protocol |
| 2 | ============ |
| 3 | |
Ian McDonald | 98069ff | 2005-11-10 13:04:33 -0800 | [diff] [blame] | 4 | |
| 5 | Contents |
| 6 | ======== |
| 7 | |
| 8 | - Introduction |
| 9 | - Missing features |
| 10 | - Socket options |
| 11 | - Notes |
| 12 | |
| 13 | Introduction |
| 14 | ============ |
| 15 | |
| 16 | Datagram Congestion Control Protocol (DCCP) is an unreliable, connection |
| 17 | based protocol designed to solve issues present in UDP and TCP particularly |
| 18 | for real time and multimedia traffic. |
| 19 | |
| 20 | It has a base protocol and pluggable congestion control IDs (CCIDs). |
| 21 | |
Ian McDonald | 5fce9a2 | 2006-12-09 23:58:10 -0200 | [diff] [blame] | 22 | It is at proposed standard RFC status and the homepage for DCCP as a protocol |
| 23 | is at: |
Ian McDonald | ddfe10b | 2006-11-20 18:42:45 -0200 | [diff] [blame] | 24 | http://www.read.cs.ucla.edu/dccp/ |
Ian McDonald | 98069ff | 2005-11-10 13:04:33 -0800 | [diff] [blame] | 25 | |
| 26 | Missing features |
| 27 | ================ |
| 28 | |
| 29 | The DCCP implementation does not currently have all the features that are in |
Ian McDonald | ddfe10b | 2006-11-20 18:42:45 -0200 | [diff] [blame] | 30 | the RFC. |
Ian McDonald | 98069ff | 2005-11-10 13:04:33 -0800 | [diff] [blame] | 31 | |
Ian McDonald | ddfe10b | 2006-11-20 18:42:45 -0200 | [diff] [blame] | 32 | The known bugs are at: |
| 33 | http://linux-net.osdl.org/index.php/TODO#DCCP |
Ian McDonald | 98069ff | 2005-11-10 13:04:33 -0800 | [diff] [blame] | 34 | |
| 35 | Socket options |
| 36 | ============== |
| 37 | |
Gerrit Renker | 00e4d11 | 2006-09-22 09:33:58 +0100 | [diff] [blame] | 38 | DCCP_SOCKOPT_SERVICE sets the service. The specification mandates use of |
| 39 | service codes (RFC 4340, sec. 8.1.2); if this socket option is not set, |
| 40 | the socket will fall back to 0 (which means that no meaningful service code |
| 41 | is present). Connecting sockets set at most one service option; for |
| 42 | listening sockets, multiple service codes can be specified. |
Ian McDonald | 98069ff | 2005-11-10 13:04:33 -0800 | [diff] [blame] | 43 | |
Gerrit Renker | 6f4e5ff | 2006-11-10 17:43:06 -0200 | [diff] [blame] | 44 | DCCP_SOCKOPT_SEND_CSCOV and DCCP_SOCKOPT_RECV_CSCOV are used for setting the |
| 45 | partial checksum coverage (RFC 4340, sec. 9.2). The default is that checksums |
| 46 | always cover the entire packet and that only fully covered application data is |
| 47 | accepted by the receiver. Hence, when using this feature on the sender, it must |
| 48 | be enabled at the receiver, too with suitable choice of CsCov. |
| 49 | |
| 50 | DCCP_SOCKOPT_SEND_CSCOV sets the sender checksum coverage. Values in the |
| 51 | range 0..15 are acceptable. The default setting is 0 (full coverage), |
| 52 | values between 1..15 indicate partial coverage. |
| 53 | DCCP_SOCKOPT_SEND_CSCOV is for the receiver and has a different meaning: it |
| 54 | sets a threshold, where again values 0..15 are acceptable. The default |
| 55 | of 0 means that all packets with a partial coverage will be discarded. |
| 56 | Values in the range 1..15 indicate that packets with minimally such a |
| 57 | coverage value are also acceptable. The higher the number, the more |
| 58 | restrictive this setting (see [RFC 4340, sec. 9.2.1]). |
| 59 | |
Gerrit Renker | 2e2e9e9 | 2006-11-13 13:23:52 -0200 | [diff] [blame] | 60 | Sysctl variables |
| 61 | ================ |
| 62 | Several DCCP default parameters can be managed by the following sysctls |
| 63 | (sysctl net.dccp.default or /proc/sys/net/dccp/default): |
| 64 | |
| 65 | request_retries |
| 66 | The number of active connection initiation retries (the number of |
| 67 | Requests minus one) before timing out. In addition, it also governs |
| 68 | the behaviour of the other, passive side: this variable also sets |
| 69 | the number of times DCCP repeats sending a Response when the initial |
| 70 | handshake does not progress from RESPOND to OPEN (i.e. when no Ack |
| 71 | is received after the initial Request). This value should be greater |
| 72 | than 0, suggested is less than 10. Analogue of tcp_syn_retries. |
| 73 | |
| 74 | retries1 |
| 75 | How often a DCCP Response is retransmitted until the listening DCCP |
| 76 | side considers its connecting peer dead. Analogue of tcp_retries1. |
| 77 | |
| 78 | retries2 |
| 79 | The number of times a general DCCP packet is retransmitted. This has |
| 80 | importance for retransmitted acknowledgments and feature negotiation, |
| 81 | data packets are never retransmitted. Analogue of tcp_retries2. |
| 82 | |
| 83 | send_ndp = 1 |
| 84 | Whether or not to send NDP count options (sec. 7.7.2). |
| 85 | |
| 86 | send_ackvec = 1 |
| 87 | Whether or not to send Ack Vector options (sec. 11.5). |
| 88 | |
| 89 | ack_ratio = 2 |
| 90 | The default Ack Ratio (sec. 11.3) to use. |
| 91 | |
| 92 | tx_ccid = 2 |
| 93 | Default CCID for the sender-receiver half-connection. |
| 94 | |
| 95 | rx_ccid = 2 |
| 96 | Default CCID for the receiver-sender half-connection. |
| 97 | |
| 98 | seq_window = 100 |
| 99 | The initial sequence window (sec. 7.5.2). |
| 100 | |
Ian McDonald | 82e3ab9 | 2006-11-20 19:19:32 -0200 | [diff] [blame] | 101 | tx_qlen = 5 |
| 102 | The size of the transmit buffer in packets. A value of 0 corresponds |
| 103 | to an unbounded transmit buffer. |
| 104 | |
Ian McDonald | 98069ff | 2005-11-10 13:04:33 -0800 | [diff] [blame] | 105 | Notes |
| 106 | ===== |
| 107 | |
Ian McDonald | ddfe10b | 2006-11-20 18:42:45 -0200 | [diff] [blame] | 108 | DCCP does not travel through NAT successfully at present on many boxes. This is |
| 109 | because the checksum covers the psuedo-header as per TCP and UDP. Linux NAT |
| 110 | support for DCCP has been added. |