tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
Documentation: add more details in tipc.rst
kernel-doc for TIPC is too simple, we need to add more information for it.
This patch is to extend the abstract, and add the Features and Links items.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
authored by
Xin Long
and committed by
David S. Miller
09ef1786
4f408e1f
+118
-3
+118
-3
Documentation/networking/tipc.rst
+118
-3
Documentation/networking/tipc.rst
···
4
4
Linux Kernel TIPC
5
5
=================
6
6
7
-
TIPC (Transparent Inter Process Communication) is a protocol that is
8
-
specially designed for intra-cluster communication.
7
+
Introduction
8
+
============
9
9
10
-
For more information about TIPC, see http://tipc.sourceforge.net.
10
+
TIPC (Transparent Inter Process Communication) is a protocol that is specially
11
+
designed for intra-cluster communication. It can be configured to transmit
12
+
messages either on UDP or directly across Ethernet. Message delivery is
13
+
sequence guaranteed, loss free and flow controlled. Latency times are shorter
14
+
than with any other known protocol, while maximal throughput is comparable to
15
+
that of TCP.
16
+
17
+
TIPC Features
18
+
-------------
19
+
20
+
- Cluster wide IPC service
21
+
22
+
Have you ever wished you had the convenience of Unix Domain Sockets even when
23
+
transmitting data between cluster nodes? Where you yourself determine the
24
+
addresses you want to bind to and use? Where you don't have to perform DNS
25
+
lookups and worry about IP addresses? Where you don't have to start timers
26
+
to monitor the continuous existence of peer sockets? And yet without the
27
+
downsides of that socket type, such as the risk of lingering inodes?
28
+
29
+
Welcome to the Transparent Inter Process Communication service, TIPC in short,
30
+
which gives you all of this, and a lot more.
31
+
32
+
- Service Addressing
33
+
34
+
A fundamental concept in TIPC is that of Service Addressing which makes it
35
+
possible for a programmer to chose his own address, bind it to a server
36
+
socket and let client programs use only that address for sending messages.
37
+
38
+
- Service Tracking
39
+
40
+
A client wanting to wait for the availability of a server, uses the Service
41
+
Tracking mechanism to subscribe for binding and unbinding/close events for
42
+
sockets with the associated service address.
43
+
44
+
The service tracking mechanism can also be used for Cluster Topology Tracking,
45
+
i.e., subscribing for availability/non-availability of cluster nodes.
46
+
47
+
Likewise, the service tracking mechanism can be used for Cluster Connectivity
48
+
Tracking, i.e., subscribing for up/down events for individual links between
49
+
cluster nodes.
50
+
51
+
- Transmission Modes
52
+
53
+
Using a service address, a client can send datagram messages to a server socket.
54
+
55
+
Using the same address type, it can establish a connection towards an accepting
56
+
server socket.
57
+
58
+
It can also use a service address to create and join a Communication Group,
59
+
which is the TIPC manifestation of a brokerless message bus.
60
+
61
+
Multicast with very good performance and scalability is available both in
62
+
datagram mode and in communication group mode.
63
+
64
+
- Inter Node Links
65
+
66
+
Communication between any two nodes in a cluster is maintained by one or two
67
+
Inter Node Links, which both guarantee data traffic integrity and monitor
68
+
the peer node's availability.
69
+
70
+
- Cluster Scalability
71
+
72
+
By applying the Overlapping Ring Monitoring algorithm on the inter node links
73
+
it is possible to scale TIPC clusters up to 1000 nodes with a maintained
74
+
neighbor failure discovery time of 1-2 seconds. For smaller clusters this
75
+
time can be made much shorter.
76
+
77
+
- Neighbor Discovery
78
+
79
+
Neighbor Node Discovery in the cluster is done by Ethernet broadcast or UDP
80
+
multicast, when any of those services are available. If not, configured peer
81
+
IP addresses can be used.
82
+
83
+
- Configuration
84
+
85
+
When running TIPC in single node mode no configuration whatsoever is needed.
86
+
When running in cluster mode TIPC must as a minimum be given a node address
87
+
(before Linux 4.17) and told which interface to attach to. The "tipc"
88
+
configuration tool makes is possible to add and maintain many more
89
+
configuration parameters.
90
+
91
+
- Performance
92
+
93
+
TIPC message transfer latency times are better than in any other known protocol.
94
+
Maximal byte throughput for inter-node connections is still somewhat lower than
95
+
for TCP, while they are superior for intra-node and inter-container throughput
96
+
on the same host.
97
+
98
+
- Language Support
99
+
100
+
The TIPC user API has support for C, Python, Perl, Ruby, D and Go.
101
+
102
+
More Information
103
+
----------------
104
+
105
+
- How to set up TIPC:
106
+
107
+
http://tipc.io/getting_started.html
108
+
109
+
- How to program with TIPC:
110
+
111
+
http://tipc.io/programming.html
112
+
113
+
- How to contribute to TIPC:
114
+
115
+
- http://tipc.io/contacts.html
116
+
117
+
- More details about TIPC specification:
118
+
119
+
http://tipc.io/protocol.html
120
+
121
+
122
+
Implementation
123
+
==============
124
+
125
+
TIPC is implemented as a kernel module in net/tipc/ directory.
11
126
12
127
TIPC Base Types
13
128
---------------