JGroups的协议栈

JGroups的协议栈

在组通信领域，JGroups作为可靠的多播通信工具集享有盛名。JGroups的一大特点就是其灵活的协议栈 Protocol stack, 灵活固然好，有时不明就里，胡乱搭配，反而不妙。

所以，要搞清楚协议的应用场景和参数配置意义，举例如下

< config >
      < UDP
          mcast_port ="${jgroups.udp.mcast_port:45588}"
         tos ="8"
         ucast_recv_buf_size ="20M"
         ucast_send_buf_size ="640K"
         mcast_recv_buf_size ="25M"
         mcast_send_buf_size ="640K"
         loopback ="true"
         discard_incompatible_packets ="true"
         max_bundle_size ="64K"
         max_bundle_timeout ="30"
         ip_ttl ="${jgroups.udp.ip_ttl:2}"
         enable_bundling ="true"
         enable_diagnostics ="true"
         thread_naming_pattern ="cl"
         timer.num_threads ="4"

         thread_pool.enabled ="true"
         thread_pool.min_threads ="2"
         thread_pool.max_threads ="8"
         thread_pool.keep_alive_time ="5000"
         thread_pool.queue_enabled ="true"
         thread_pool.queue_max_size ="10000"
         thread_pool.rejection_policy ="discard"

         oob_thread_pool.enabled ="true"
         oob_thread_pool.min_threads ="1"
         oob_thread_pool.max_threads ="8"
         oob_thread_pool.keep_alive_time ="5000"
         oob_thread_pool.queue_enabled ="false"
         oob_thread_pool.queue_max_size ="100"
         oob_thread_pool.rejection_policy ="Run" />

     < PING  timeout ="2000"
            num_initial_members ="3" />
     < MERGE2  max_interval ="30000"
            min_interval ="10000" />
     < FD_SOCK />
     < FD_ALL />
     < VERIFY_SUSPECT  timeout ="1500"    />
     < BARRIER  />
     < pbcast.NAKACK  use_stats_for_retransmission ="false"
                   exponential_backoff ="0"
                   use_mcast_xmit ="true"  gc_lag ="0"
                   retransmit_timeout ="300,600,1200"
                   discard_delivered_msgs ="true" />
     < UNICAST  timeout ="300,600,1200" />
     < pbcast.STABLE  stability_delay ="1000"  desired_avg_gossip ="50000"
                   max_bytes ="1M" />
     < pbcast.GMS  print_local_addr ="false"  join_timeout ="3000"

                view_bundling ="true" />
     < FC  max_credits ="500K"
                    min_threshold ="0.20" />
     < FRAG2  frag_size ="60K"    />
     < pbcast.STREAMING_STATE_TRANSFER  />
     <!-- pbcast.STATE_TRANSFER  / -->
     <!--  pbcast.FLUSH  / -->
</ config >

 

1. UDP 协议，学过计算机网络的人都知道UDP是传输层的基本协议，特点就是快但不一定可靠， 这里是用基于UDP的多播multicast协议来发送消息的， 也可以改成其他传输协议，如 tcp, tcp_nio, tunnel 

This is the transport protocol. It uses IP multicasting to send messages to the entire cluster, or individual nodes. Other transports include TCP, TCP_NIO and TUNNEL.

2. PING协议， 用缺省的IP多播来发现最初的成员， 类似于操作系统的 ping 命令，一旦ping通，则一个单播join请求将被发送至它以加入当前的 group 或者叫做cluster 

Uses IP multicast (by default) to find initial members. Once found, the current coordinator can be determined and a unicast JOIN request will be sent to it in order to join the cluster.

3. MERGE2, 用来合并子组，或在一个子组分离时去除子组 

Will merge subgroups back into one group, kicks in after a cluster partition.

4. FD_SOCK 基于socket 的错误检测协议 (在成员之间形成的一个环)，当某个成员失败时会生成通知 

Failure detection based on sockets (in a ring form between members). Generates notification if a member fails

5. FD 基于心跳 heartbeat 和 你还活着 are-you-live 消息的错误检测协议 (在成员之间形成的一个环)，当某个成员失败时会生成通知

Failure detection based on heartbeats and are-you-alive messages (in a ring  form between members). Generates notification if a member fails

6. VERIFY_SUSPECT 重复检测一个挂起的成员 是否 已经死掉了， 否则以下协议产生的消息将被丢弃 

Double-checks whether a suspected member is really dead, otherwise the suspicion generated from protocol below is discarded

7. pbcast.NAKACK 确保消息的可靠和 FIFO 顺序 , 消息可靠性保证了一个消息会被收到，否则接收者会要求重传，FIFO保证了从发送者发过来的所有消息会以相同顺序到达接收方 

Ensures (a) message reliability and (b) FIFO. Message reliability guarantees that a message will be received. If not, the receiver(s) will request retransmission. FIFO guarantees that all messages from sender P will be received in the order P sent them

8. UNICAST 与NAKACK差不多，只不过它不是多播，而是单播的 

Same as NAKACK for unicast messages: messages from sender P will not be lost (retransmission if necessary) and will be in FIFO order (conceptually the same as TCP in TCP/IP)

9. pbcast.STABLE 将已被所有成员收到的消息删除，即分布式垃圾回收 

Deletes messages that have been seen by all members (distributed message garbage collection)

10. pbcast.GMS 成员协议，负责成员的 加入/离开 并更新新的成员视图 

Membership protocol. Responsible for joining/leaving members and installing new views.

11. FRAG2 将巨大的消息打碎分割成若干个小的发送，并在接收端重新组装回去，适用于多播和单播协议 

Fragments large messages into smaller ones and reassembles them back at the receiver side. For both multicast and unicast messages

转自：http://www.cnblogs.com/walterfan/archive/2011/03/31/2000526.html mark下