其他分享
首页 > 其他分享> > 实验6:开源控制器实践——RYU

实验6:开源控制器实践——RYU

作者:互联网

SDN实验6:开源控制器实践RYU

实验目的

基本要求

在Ryu安装目录下执行ryu --version查看版本

建立拓扑:

sudo mn --topo=single,3 --mac --controller=remote,ip=127.0.0.1,port=6633 --switch ovsk,protocols=OpenFlow10

连接ryu控制器:

ryu-manager ryu/ryu/app/gui_topology/gui_topology.py --observe-links

L2Switch.py文件:

from ryu.base import app_manager
from ryu.controller import ofp_event
from ryu.controller.handler import MAIN_DISPATCHER
from ryu.controller.handler import set_ev_cls
from ryu.ofproto import ofproto_v1_0

class L2Switch(app_manager.RyuApp):
    OFP_VERSIONS = [ofproto_v1_0.OFP_VERSION]

    def __init__(self, *args, **kwargs):
        super(L2Switch, self).__init__(*args, **kwargs)

    @set_ev_cls(ofp_event.EventOFPPacketIn, MAIN_DISPATCHER)
    def packet_in_handler(self, ev):
        msg = ev.msg
        dp = msg.datapath
        ofp = dp.ofproto
        ofp_parser = dp.ofproto_parser

        actions = [ofp_parser.OFPActionOutput(ofp.OFPP_FLOOD)]

        data = None
        if msg.buffer_id == ofp.OFP_NO_BUFFER:
             data = msg.data

        out = ofp_parser.OFPPacketOut(
            datapath=dp, buffer_id=msg.buffer_id, in_port=msg.in_port,
            actions=actions, data = data)
        dp.send_msg(out)

使用ryu-manager L2Switch.py运行

h1 ping h2:

h1 ping h3:

可以看到均为洪泛转发ICMP报文

不同之处在于L2Switch下发的流表无法查看,而Hub可以查看

进阶要求

# 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.

# 引入包
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):
    # 定义openflow版本
    OFP_VERSIONS = [ofproto_v1_3.OFP_VERSION]

    def __init__(self, *args, **kwargs):
        super(SimpleSwitch13, self).__init__(*args, **kwargs)
        # 定义保存mac地址到端口的一个映射
        self.mac_to_port = {}

    # 处理EventOFPSwitchFeatures事件
    @set_ev_cls(ofp_event.EventOFPSwitchFeatures, CONFIG_DISPATCHER)
    def switch_features_handler(self, ev):
        datapath = ev.msg.datapath
        ofproto = datapath.ofproto
        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()#match:流表项匹配,OFPMatch():不匹配任何信息
        actions = [parser.OFPActionOutput(ofproto.OFPP_CONTROLLER,
                                          ofproto.OFPCML_NO_BUFFER)]
        self.add_flow(datapath, 0, match, actions)#添加流表项

    # 添加流表函数
    def add_flow(self, datapath, priority, match, actions, buffer_id=None):
        # 获取交换机信息
        ofproto = datapath.ofproto
        parser = datapath.ofproto_parser

        # 对action进行包装
        inst = [parser.OFPInstructionActions(ofproto.OFPIT_APPLY_ACTIONS,
                                             actions)]
        # 判断是否有buffer_id,生成mod对象
        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)
        # 发送mod
        datapath.send_msg(mod)

    # 触发packet in事件时,调用_packet_in_handler函数
    @set_ev_cls(ofp_event.EventOFPPacketIn, MAIN_DISPATCHER)
    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)
        # 获取包信息,交换机信息,协议等等
        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]

        # 忽略LLDP类型
        if eth.ethertype == ether_types.ETH_TYPE_LLDP:
            # ignore lldp packet
            return

        # 获取源端口,目的端口
        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)

        # 学习包的源地址,和交换机上的入端口绑定
        # learn a mac address to avoid FLOOD next time.
        self.mac_to_port[dpid][src] = in_port

        # 查看是否已经学习过该目的mac地址
        if dst in self.mac_to_port[dpid]:
            out_port = self.mac_to_port[dpid][dst]
        # 否则进行洪泛
        else:
            out_port = ofproto.OFPP_FLOOD

        actions = [parser.OFPActionOutput(out_port)]

        # 下发流表处理后续包,不再触发 packet in 事件
        # 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:
                self.add_flow(datapath, 1, match, actions, msg.buffer_id)
                return
            else:
                self.add_flow(datapath, 1, match, actions)
        data = None
        if msg.buffer_id == ofproto.OFP_NO_BUFFER:
            data = msg.data
		# 发送Packet_out数据包
        out = parser.OFPPacketOut(datapath=datapath, buffer_id=msg.buffer_id,
                                  in_port=in_port, actions=actions, data=data)
        # 发送流表
        datapath.send_msg(out)

保存mac地址到交换机端口的映射,为交换机自学习功能提供数据结构进行mac端口的存储

dpid = datapath.id      # simple_switch的dpid赋值
dpid = format(datapath.id, "d").zfill(16)   # simple_switch_13的dpid赋值

可见在simple_switch_13中,会在前端加上0以填充至16位,simple_switch直接输出dpid

实现了交换机以特性应答消息响应特性请求

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。如果以太网类型是LLDP类型,则不予处理。如果不是,则获取源端口目的端口,以及交换机id,先学习源地址对应的交换机的入端口,再查看是否已经学习目的mac地址,如果没有则进行洪泛转发。如果学习过该mac地址,则查看是否有buffer_id,如果有的话,则在添加流动作时加上buffer_id,向交换机发送流表

switch_features_handler下发流表的优先级更高

个人总结

本次实验的基础要求和实验五比较相似,虽然遇到了一些无法ping通的小bug,但重新尝试之后就消失了(所以没有搞清楚为什么会ping不通)进阶要求当中,阅读源码的部分应该是比较复杂的,需要对RYU的原理有更深入的了解,对源码分析和对openflow协议的理解也有一定的要求

标签:控制器,ryu,datapath,id,ofproto,开源,msg,RYU,port
来源: https://www.cnblogs.com/CCchaos/p/15394271.html