2 # IP Virtual Server configuration
5 tristate "IP virtual server support"
6 depends on NET && INET && NETFILTER
8 IP Virtual Server support will let you build a high-performance
9 virtual server based on cluster of two or more real servers. This
10 option must be enabled for at least one of the clustered computers
11 that will take care of intercepting incoming connections to a
12 single IP address and scheduling them to real servers.
14 Three request dispatching techniques are implemented, they are
15 virtual server via NAT, virtual server via tunneling and virtual
16 server via direct routing. The several scheduling algorithms can
17 be used to choose which server the connection is directed to,
18 thus load balancing can be achieved among the servers. For more
19 information and its administration program, please visit the
20 following URL: <http://www.linuxvirtualserver.org/>.
22 If you want to compile it in kernel, say Y. To compile it as a
23 module, choose M here. If unsure, say N.
28 bool "IPv6 support for IPVS"
29 depends on EXPERIMENTAL && (IPV6 = y || IP_VS = IPV6)
31 Add IPv6 support to IPVS. This is incomplete and might be dangerous.
33 See http://www.mindbasket.com/ipvs for more information.
38 bool "IP virtual server debugging"
40 Say Y here if you want to get additional messages useful in
41 debugging the IP virtual server code. You can change the debug
42 level in /proc/sys/net/ipv4/vs/debug_level
45 int "IPVS connection table size (the Nth power of 2)"
49 The IPVS connection hash table uses the chaining scheme to handle
50 hash collisions. Using a big IPVS connection hash table will greatly
51 reduce conflicts when there are hundreds of thousands of connections
54 Note the table size must be power of 2. The table size will be the
55 value of 2 to the your input number power. The number to choose is
56 from 8 to 20, the default number is 12, which means the table size
57 is 4096. Don't input the number too small, otherwise you will lose
58 performance on it. You can adapt the table size yourself, according
59 to your virtual server application. It is good to set the table size
60 not far less than the number of connections per second multiplying
61 average lasting time of connection in the table. For example, your
62 virtual server gets 200 connections per second, the connection lasts
63 for 200 seconds in average in the connection table, the table size
64 should be not far less than 200x200, it is good to set the table
67 Another note that each connection occupies 128 bytes effectively and
68 each hash entry uses 8 bytes, so you can estimate how much memory is
71 You can overwrite this number setting conn_tab_bits module parameter
72 or by appending ip_vs.conn_tab_bits=? to the kernel command line
73 if IP VS was compiled built-in.
75 comment "IPVS transport protocol load balancing support"
77 config IP_VS_PROTO_TCP
78 bool "TCP load balancing support"
80 This option enables support for load balancing TCP transport
81 protocol. Say Y if unsure.
83 config IP_VS_PROTO_UDP
84 bool "UDP load balancing support"
86 This option enables support for load balancing UDP transport
87 protocol. Say Y if unsure.
89 config IP_VS_PROTO_AH_ESP
93 config IP_VS_PROTO_ESP
94 bool "ESP load balancing support"
95 select IP_VS_PROTO_AH_ESP
97 This option enables support for load balancing ESP (Encapsulation
98 Security Payload) transport protocol. Say Y if unsure.
100 config IP_VS_PROTO_AH
101 bool "AH load balancing support"
102 select IP_VS_PROTO_AH_ESP
104 This option enables support for load balancing AH (Authentication
105 Header) transport protocol. Say Y if unsure.
107 config IP_VS_PROTO_SCTP
108 bool "SCTP load balancing support"
111 This option enables support for load balancing SCTP transport
112 protocol. Say Y if unsure.
114 comment "IPVS scheduler"
117 tristate "round-robin scheduling"
119 The robin-robin scheduling algorithm simply directs network
120 connections to different real servers in a round-robin manner.
122 If you want to compile it in kernel, say Y. To compile it as a
123 module, choose M here. If unsure, say N.
126 tristate "weighted round-robin scheduling"
129 The weighted robin-robin scheduling algorithm directs network
130 connections to different real servers based on server weights
131 in a round-robin manner. Servers with higher weights receive
132 new connections first than those with less weights, and servers
133 with higher weights get more connections than those with less
134 weights and servers with equal weights get equal connections.
136 If you want to compile it in kernel, say Y. To compile it as a
137 module, choose M here. If unsure, say N.
140 tristate "least-connection scheduling"
142 The least-connection scheduling algorithm directs network
143 connections to the server with the least number of active
146 If you want to compile it in kernel, say Y. To compile it as a
147 module, choose M here. If unsure, say N.
150 tristate "weighted least-connection scheduling"
152 The weighted least-connection scheduling algorithm directs network
153 connections to the server with the least active connections
154 normalized by the server weight.
156 If you want to compile it in kernel, say Y. To compile it as a
157 module, choose M here. If unsure, say N.
160 tristate "locality-based least-connection scheduling"
162 The locality-based least-connection scheduling algorithm is for
163 destination IP load balancing. It is usually used in cache cluster.
164 This algorithm usually directs packet destined for an IP address to
165 its server if the server is alive and under load. If the server is
166 overloaded (its active connection numbers is larger than its weight)
167 and there is a server in its half load, then allocate the weighted
168 least-connection server to this IP address.
170 If you want to compile it in kernel, say Y. To compile it as a
171 module, choose M here. If unsure, say N.
174 tristate "locality-based least-connection with replication scheduling"
176 The locality-based least-connection with replication scheduling
177 algorithm is also for destination IP load balancing. It is
178 usually used in cache cluster. It differs from the LBLC scheduling
179 as follows: the load balancer maintains mappings from a target
180 to a set of server nodes that can serve the target. Requests for
181 a target are assigned to the least-connection node in the target's
182 server set. If all the node in the server set are over loaded,
183 it picks up a least-connection node in the cluster and adds it
184 in the sever set for the target. If the server set has not been
185 modified for the specified time, the most loaded node is removed
186 from the server set, in order to avoid high degree of replication.
188 If you want to compile it in kernel, say Y. To compile it as a
189 module, choose M here. If unsure, say N.
192 tristate "destination hashing scheduling"
194 The destination hashing scheduling algorithm assigns network
195 connections to the servers through looking up a statically assigned
196 hash table by their destination IP addresses.
198 If you want to compile it in kernel, say Y. To compile it as a
199 module, choose M here. If unsure, say N.
202 tristate "source hashing scheduling"
204 The source hashing scheduling algorithm assigns network
205 connections to the servers through looking up a statically assigned
206 hash table by their source IP addresses.
208 If you want to compile it in kernel, say Y. To compile it as a
209 module, choose M here. If unsure, say N.
212 tristate "shortest expected delay scheduling"
214 The shortest expected delay scheduling algorithm assigns network
215 connections to the server with the shortest expected delay. The
216 expected delay that the job will experience is (Ci + 1) / Ui if
217 sent to the ith server, in which Ci is the number of connections
218 on the ith server and Ui is the fixed service rate (weight)
221 If you want to compile it in kernel, say Y. To compile it as a
222 module, choose M here. If unsure, say N.
225 tristate "never queue scheduling"
227 The never queue scheduling algorithm adopts a two-speed model.
228 When there is an idle server available, the job will be sent to
229 the idle server, instead of waiting for a fast one. When there
230 is no idle server available, the job will be sent to the server
231 that minimize its expected delay (The Shortest Expected Delay
232 scheduling algorithm).
234 If you want to compile it in kernel, say Y. To compile it as a
235 module, choose M here. If unsure, say N.
237 comment 'IPVS application helper'
240 tristate "FTP protocol helper"
241 depends on IP_VS_PROTO_TCP
243 FTP is a protocol that transfers IP address and/or port number in
244 the payload. In the virtual server via Network Address Translation,
245 the IP address and port number of real servers cannot be sent to
246 clients in ftp connections directly, so FTP protocol helper is
247 required for tracking the connection and mangling it back to that of
250 If you want to compile it in kernel, say Y. To compile it as a
251 module, choose M here. If unsure, say N.