2 # IP Virtual Server configuration
5 tristate "IP virtual server support (EXPERIMENTAL)"
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 "IP virtual server debugging"
30 Say Y here if you want to get additional messages useful in
31 debugging the IP virtual server code. You can change the debug
32 level in /proc/sys/net/ipv4/vs/debug_level
35 int "IPVS connection table size (the Nth power of 2)"
38 The IPVS connection hash table uses the chaining scheme to handle
39 hash collisions. Using a big IPVS connection hash table will greatly
40 reduce conflicts when there are hundreds of thousands of connections
43 Note the table size must be power of 2. The table size will be the
44 value of 2 to the your input number power. The number to choose is
45 from 8 to 20, the default number is 12, which means the table size
46 is 4096. Don't input the number too small, otherwise you will lose
47 performance on it. You can adapt the table size yourself, according
48 to your virtual server application. It is good to set the table size
49 not far less than the number of connections per second multiplying
50 average lasting time of connection in the table. For example, your
51 virtual server gets 200 connections per second, the connection lasts
52 for 200 seconds in average in the connection table, the table size
53 should be not far less than 200x200, it is good to set the table
56 Another note that each connection occupies 128 bytes effectively and
57 each hash entry uses 8 bytes, so you can estimate how much memory is
60 comment "IPVS transport protocol load balancing support"
62 config IP_VS_PROTO_TCP
63 bool "TCP load balancing support"
65 This option enables support for load balancing TCP transport
66 protocol. Say Y if unsure.
68 config IP_VS_PROTO_UDP
69 bool "UDP load balancing support"
71 This option enables support for load balancing UDP transport
72 protocol. Say Y if unsure.
74 config IP_VS_PROTO_AH_ESP
78 config IP_VS_PROTO_ESP
79 bool "ESP load balancing support"
80 select IP_VS_PROTO_AH_ESP
82 This option enables support for load balancing ESP (Encapsulation
83 Security Payload) transport protocol. Say Y if unsure.
86 bool "AH load balancing support"
87 select IP_VS_PROTO_AH_ESP
89 This option enables support for load balancing AH (Authentication
90 Header) transport protocol. Say Y if unsure.
92 comment "IPVS scheduler"
95 tristate "round-robin scheduling"
97 The robin-robin scheduling algorithm simply directs network
98 connections to different real servers in a round-robin manner.
100 If you want to compile it in kernel, say Y. To compile it as a
101 module, choose M here. If unsure, say N.
104 tristate "weighted round-robin scheduling"
106 The weighted robin-robin scheduling algorithm directs network
107 connections to different real servers based on server weights
108 in a round-robin manner. Servers with higher weights receive
109 new connections first than those with less weights, and servers
110 with higher weights get more connections than those with less
111 weights and servers with equal weights get equal connections.
113 If you want to compile it in kernel, say Y. To compile it as a
114 module, choose M here. If unsure, say N.
117 tristate "least-connection scheduling"
119 The least-connection scheduling algorithm directs network
120 connections to the server with the least number of active
123 If you want to compile it in kernel, say Y. To compile it as a
124 module, choose M here. If unsure, say N.
127 tristate "weighted least-connection scheduling"
129 The weighted least-connection scheduling algorithm directs network
130 connections to the server with the least active connections
131 normalized by the server weight.
133 If you want to compile it in kernel, say Y. To compile it as a
134 module, choose M here. If unsure, say N.
137 tristate "locality-based least-connection scheduling"
139 The locality-based least-connection scheduling algorithm is for
140 destination IP load balancing. It is usually used in cache cluster.
141 This algorithm usually directs packet destined for an IP address to
142 its server if the server is alive and under load. If the server is
143 overloaded (its active connection numbers is larger than its weight)
144 and there is a server in its half load, then allocate the weighted
145 least-connection server to this IP address.
147 If you want to compile it in kernel, say Y. To compile it as a
148 module, choose M here. If unsure, say N.
151 tristate "locality-based least-connection with replication scheduling"
153 The locality-based least-connection with replication scheduling
154 algorithm is also for destination IP load balancing. It is
155 usually used in cache cluster. It differs from the LBLC scheduling
156 as follows: the load balancer maintains mappings from a target
157 to a set of server nodes that can serve the target. Requests for
158 a target are assigned to the least-connection node in the target's
159 server set. If all the node in the server set are over loaded,
160 it picks up a least-connection node in the cluster and adds it
161 in the sever set for the target. If the server set has not been
162 modified for the specified time, the most loaded node is removed
163 from the server set, in order to avoid high degree of replication.
165 If you want to compile it in kernel, say Y. To compile it as a
166 module, choose M here. If unsure, say N.
169 tristate "destination hashing scheduling"
171 The destination hashing scheduling algorithm assigns network
172 connections to the servers through looking up a statically assigned
173 hash table by their destination IP addresses.
175 If you want to compile it in kernel, say Y. To compile it as a
176 module, choose M here. If unsure, say N.
179 tristate "source hashing scheduling"
181 The source hashing scheduling algorithm assigns network
182 connections to the servers through looking up a statically assigned
183 hash table by their source IP addresses.
185 If you want to compile it in kernel, say Y. To compile it as a
186 module, choose M here. If unsure, say N.
189 tristate "shortest expected delay scheduling"
191 The shortest expected delay scheduling algorithm assigns network
192 connections to the server with the shortest expected delay. The
193 expected delay that the job will experience is (Ci + 1) / Ui if
194 sent to the ith server, in which Ci is the number of connections
195 on the ith server and Ui is the fixed service rate (weight)
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 "never queue scheduling"
204 The never queue scheduling algorithm adopts a two-speed model.
205 When there is an idle server available, the job will be sent to
206 the idle server, instead of waiting for a fast one. When there
207 is no idle server available, the job will be sent to the server
208 that minimize its expected delay (The Shortest Expected Delay
209 scheduling algorithm).
211 If you want to compile it in kernel, say Y. To compile it as a
212 module, choose M here. If unsure, say N.
214 comment 'IPVS application helper'
217 tristate "FTP protocol helper"
218 depends on IP_VS_PROTO_TCP
220 FTP is a protocol that transfers IP address and/or port number in
221 the payload. In the virtual server via Network Address Translation,
222 the IP address and port number of real servers cannot be sent to
223 clients in ftp connections directly, so FTP protocol helper is
224 required for tracking the connection and mangling it back to that of
227 If you want to compile it in kernel, say Y. To compile it as a
228 module, choose M here. If unsure, say N.