实验6:开源控制器实践——RYU

实验6:开源控制器实践——RYU

一、实验目的

  1. 能够独立部署RYU控制器;
  2. 能够理解RYU控制器实现软件定义的集线器原理;
  3. 能够理解RYU控制器实现软件定义的交换机原理。

二、实验环境

Ubuntu 20.04 Desktop amd64

三、实验要求

(一)基本要求

  1. 搭建下图所示SDN拓扑,协议使用Open Flow 1.0,并连接Ryu控制器,通过Ryu的图形界面查看网络拓扑。

搭建如上图所示的拓扑,并开启ryu控制器

 

搭建如上图所示的拓扑,并开启ryu控制器

 

通过Ryu的图形界面查看网络拓扑

 

 

L2Switch模块代码

 

 

h1 ping h2(观察到h2,h3都接收到数据包)

 

 

 h1 ping h3(观察到h2,h3都接收到数据包)

 

 

通过dpctl dump-flows命令检查ryu的L2Switch模块和pox的Hub模块的区别

在连接ryu控制器,启动L2Switch模块后查看下发流表

 

 

在连接pox控制器,启动Hub模块后查看下发流表

 

 

编程修改L2Switch.py,另存为L2xxxxxxxxx.py,使之和POX的Hub模块的变得一致?(xxxxxxxxx为学号)

 

 

 

二)进阶要求

  1. 阅读Ryu关于simple_switch.py和simple_switch_1x.py的实现,以simple_switch_13.py为例,完成其代码的注释工作,并回答下列问题:

# Copyright (C) 2011 Nippon Telegraph and Telephone Corporation.

#

# Licensed under the Apache License, Version 2.0 (the "License");

# you may not use this file except in compliance with the License.

# You may obtain a copy of the License at

#

#    http://www.apache.org/licenses/LICENSE-2.0

#

# Unless required by applicable law or agreed to in writing, software

# distributed under the License is distributed on an "AS IS" BASIS,

# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or

# implied.

# See the License for the specific language governing permissions and

# limitations under the License.

 

#导入需要使用的相应包

from ryu.base import app_manager

from ryu.controller import ofp_event

from ryu.controller.handler import CONFIG_DISPATCHER, MAIN_DISPATCHER

from ryu.controller.handler import set_ev_cls

from ryu.ofproto import ofproto_v1_3

from ryu.lib.packet import packet

from ryu.lib.packet import ethernet

from ryu.lib.packet import ether_types

 

 

class SimpleSwitch13(app_manager.RyuApp):

    #指定OpenFlow 1.3版本

    OFP_VERSIONS = [ofproto_v1_3.OFP_VERSION]

 

    def __init__(self, *args, **kwargs):

        super(SimpleSwitch13, self).__init__(*args, **kwargs)

  #self.mac_to_port是mac地址映射到转发端口的字典。

        self.mac_to_port = {}

    @set_ev_cls(ofp_event.EventOFPSwitchFeatures, CONFIG_DISPATCHER)

    def switch_features_handler(self, ev):

        # ev.msg 是用来存储对应事件的 OpenFlow 消息类别实体

        datapath = ev.msg.datapath  

  # ofproto表示使用的OpenFlow版本所对应的ryu.ofproto.ofproto_v1_3

        ofproto = datapath.ofproto 

  # 使用对应版本的ryu.ofproto.ofproto_v1_3_parser来解析协议

        parser = datapath.ofproto_parser

 

        # install table-miss flow entry

        #

        # We specify NO BUFFER to max_len of the output action due to

        # OVS bug. At this moment, if we specify a lesser number, e.g.,

        # 128, OVS will send Packet-In with invalid buffer_id and

        # truncated packet data. In that case, we cannot output packets

        # correctly.  The bug has been fixed in OVS v2.1.0.

        match = parser.OFPMatch()

        actions = [parser.OFPActionOutput(ofproto.OFPP_CONTROLLER,

                                          ofproto.OFPCML_NO_BUFFER)]

  # priority = 0表示优先级最低,即若所有流表都匹配不到时,才会把数据包发送到controller

        self.add_flow(datapath, 0, match, actions)

 

    # 执行 add_flow() 方法以发送 Flow Mod 消息

    def add_flow(self, datapath, priority, match, actions, buffer_id=None):

        ofproto = datapath.ofproto

        parser = datapath.ofproto_parser

 

        inst = [parser.OFPInstructionActions(ofproto.OFPIT_APPLY_ACTIONS,

                                             actions)]

        if buffer_id: 

            mod = parser.OFPFlowMod(datapath=datapath, buffer_id=buffer_id,

                                    priority=priority, match=match,

                                    instructions=inst)

        else:

            mod = parser.OFPFlowMod(datapath=datapath, priority=priority,

                                    match=match, instructions=inst)

        datapath.send_msg(mod)

 

 

    @set_ev_cls(ofp_event.EventOFPPacketIn, MAIN_DISPATCHER)

    # 处理PacketIn事件

    def _packet_in_handler(self, ev):

        # If you hit this you might want to increase

        # the "miss_send_length" of your switch

        if ev.msg.msg_len < ev.msg.total_len:

            self.logger.debug("packet truncated: only %s of %s bytes",

                              ev.msg.msg_len, ev.msg.total_len)

        #从事件类里取出一些参数

  msg = ev.msg

        datapath = msg.datapath

        ofproto = datapath.ofproto

        parser = datapath.ofproto_parser

        in_port = msg.match['in_port']

 

        pkt = packet.Packet(msg.data)

        eth = pkt.get_protocols(ethernet.ethernet)[0]

 

        if eth.ethertype == ether_types.ETH_TYPE_LLDP:

      #接受到了lldp包,就直接丢弃

            # ignore lldp packet

            return

        dst = eth.dst

        src = eth.src

 

        dpid = format(datapath.id, "d").zfill(16)

        self.mac_to_port.setdefault(dpid, {})

 

        self.logger.info("packet in %s %s %s %s", dpid, src, dst, in_port)

 

        #进行自学习,尽可能避免在下一次洪泛

        # learn a mac address to avoid FLOOD next time.

  #dpid是交换机的id,src是数据包的源mac地址,in_port是交换机接受到包的端口

        self.mac_to_port[dpid][src] = in_port

 

        #检验目的地址是否已经学习

  if dst in self.mac_to_port[dpid]:

  #如果已经学习到,则向交换机下发流表,并让交换机向相应端口转发包

            out_port = self.mac_to_port[dpid][dst]

        else:

  #如果还没有学习到,则无法下发流表,让交换机洪泛转发包。

            out_port = ofproto.OFPP_FLOOD

 

        actions = [parser.OFPActionOutput(out_port)]

 

        # install a flow to avoid packet_in next time

        if out_port != ofproto.OFPP_FLOOD:

            match = parser.OFPMatch(in_port=in_port, eth_dst=dst, eth_src=src)

            # verify if we have a valid buffer_id, if yes avoid to send both

            # flow_mod & packet_out

            if msg.buffer_id != ofproto.OFP_NO_BUFFER:

      #buffer_id不为None,控制器只需下发流表的命令,交换机增加了流表项后,位于缓冲区的数据包会自动转发出去。

                self.add_flow(datapath, 1, match, actions, msg.buffer_id)

                return

            else:

      #buffer_id为None,那么控制器不仅要更改交换机的流表项,

                self.add_flow(datapath, 1, match, actions)

            #还要把数据包的信息传给交换机,让交换机把数据包转发出去。

        data = None

        if msg.buffer_id == ofproto.OFP_NO_BUFFER:

            data = msg.data

 

        out = parser.OFPPacketOut(datapath=datapath, buffer_id=msg.buffer_id,

                                  in_port=in_port, actions=actions, data=data)

        datapath.send_msg(out)
a) 代码当中的mac_to_port的作用是什么?

Mac_to_port是mac地址映射到转发端口的字典,可用于交换机自学习。
b) simple_switch和simple_switch_13在dpid的输出上有何不同?

差别在于:smiple_switch直接输出dpid,而simple_switch_13则在dpid前端填充0直到满16位
c) 相比simple_switch,simple_switch_13增加的switch_feature_handler实现了什么功能?

实现了交换机以特性应答消息来响应特性请求的功能
d) simple_switch_13是如何实现流规则下发的?

在接收到packetin事件后,首先获取包学习,交换机信息,以太网信息,协议信息等。若以太网类型是LLDP类型,则不予处理。如果不是,则获取源端口的目的端口和交换机id,先学习源地址对应的交换机的入端口,再查看是否已学习目的mac地址,如果没有则进行泛洪转发。如果学习过该mac地址,则查看是否有buffer_id,如果有的话,则在添加流表信息时加上buffer_id,向交换机发送流表。
e) switch_features_handler和_packet_in_handler两个事件在发送流规则的优先级上有何不同?

Switch_features_handler下发流表的优先级比_packet_in_handler的优先级高

(三)实践心得

在本次实践中,通过阅读RYU文档并查看相关模块的源代码,了解了RYU控制器的工作原理,并比较了RYU的L2Switch模块与POX的Hub模块的异同。本次实验的基础部分难度中等,实验操作与前两次操作ODL和POX控制器差不多,因此能较为快速地完成对应步骤,而进阶部分则难度较大,尤其在阅读源码部分进度较慢。不过,虽然阅读源码的过程中有些痛苦,但在过程中,查阅相关材料,结合源码进行阅读,也使得我对RYU的控制机制有了更为形象和深入的认识。这就是我本次实验的心得。

 

 

 

 

posted @ 2022-10-28 22:03  212106672  阅读(53)  评论(0编辑  收藏  举报