ovs实践1:基础概念和gre隧道实践操作

1.
测试环境
两个虚机:nginx-1和nginx-3
centos7 系统内核3.10.0-514.el7.x86_64
ovs_version: "2.0.0"

2.
注意事项
关闭selinux
关闭NetworkManager(必须关闭,不然出莫名其妙的网络故障)
关闭firewalld

3.
安装openvswitch(可编译安装,这里使用yum安装,简单测试)

yum install openvswitch openvswitch-devel openvswitch-test openvswitch-debuginfor
systemctl enable? openvswitch
systemctl start? openvswitch

4.
创建桥前的网络情况

创建桥前

[[email protected] ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
? ? link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
? ? inet 127.0.0.1/8 scope host lo
? ? ? valid_lft forever preferred_lft forever
? ? inet6 ::1/128 scope host
? ? ? valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
? ? link/ether 00:0c:29:80:20:b2 brd ff:ff:ff:ff:ff:ff
? ? inet 192.168.32.139/24 brd 192.168.32.255 scope global dynamic ens33
? ? ? valid_lft 1721sec preferred_lft 1721sec
? ? inet6 fe80::6d0c:96f7:cbae:c576/64 scope link
? ? ? valid_lft forever preferred_lft forever
[[email protected] ~]#
[[email protected] ~]# ip r
default via 192.168.32.2 dev ens33? proto static? metric 100
192.168.32.0/24 dev ens33? proto kernel? scope link? src 192.168.32.132? metric 100
[[email protected] ~]#

5.
ovs基础知识,参考文档:https://opengers.github.io/openstack/openstack-base-use-openvswitch/
port类型说明

normal类型
操作系统中已有的网卡(物理网卡或者虚拟机中的虚拟网卡)连接到ovs桥上,
ovs会生成一个port处理这块网卡的进出数据包.这个端口类型为normal
挂载到OVS上的网卡设备不支持分配IP地址

internal
internal端口是ovs内部创建的虚拟网卡接口
每创建一个Port,OVS会自动创建一个同名接口(Interface)挂载到新创建的Port上
interface可用来配置ip

patch
多个ovs网桥可用patch port连接,类似于veth pair.
从一个Patch Port收到的数据包会被转发到另一个Patch Port
patch port总是成对出现,分别连接在两个网桥上.使用Patch连接的两个网桥跟一个网桥没什么区别

tunnel
OVS中支持添加隧道(Tunnel)端口,常见隧道技术有两种gre或vxlan.
网络之上构建一层虚拟网络,上层应用只与虚拟网络相关.

Interface
nterface是连接到Port的网络接口设备,是OVS与外部交换数据包的组件,在通常情况下,
Port和Interface是一对一的关系,只有在配置Port为 bond模式后,Port和Interface是一对多的关系.

6.
实际操作

虚机1创建桥

[[email protected] ~]# ovs-vsctl add-br br0
[[email protected] ~]#

[[email protected] ~]# ovs-vsctl show
67ccf09b-d0d9-4ccb-9f2f-1ac7918e5c46
? ? Bridge "br0"
? ? ? ? Port "br0"
? ? ? ? ? ? Interface "br0"
? ? ? ? ? ? ? ? type: internal
? ? ovs_version: "2.0.0"
[[email protected] ~]#

虚机2创建桥

[[email protected] ~]# ovs-vsctl add-br br0
[[email protected] ~]#

[[email protected] ~]# ovs-vsctl show
25cdb171-59ee-4eb4-a8af-cd93dc460926
? ? Bridge "br0"
? ? ? ? Port "br0"
? ? ? ? ? ? Interface "br0"
? ? ? ? ? ? ? ? type: internal
? ? ovs_version: "2.0.0"

桥默认有一个br0端口和br0接口,type类型为internel
internal端口是ovs内部创建的虚拟网卡接口
每创建一个桥自动创建个类型为internal的和桥一样名字的port和interface.

7.
测试思路

虚机1和虚机2的br0各定义同网段ip,测试互通.
虚机1配置br0为192.168.100.2/24
虚机2配置br0为192.168.100.3/24

临时配置命令,见下:

[[email protected] ~]# ifconfig br0 192.168.100.2/24 up

[[email protected] ~]# ifconfig br0 192.168.100.3/24 up

在没有建立隧道前,网络是不通的.
虚机1

[[email protected] ~]# ip a |grep br0
5: br0: <BROADCAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN qlen 1000
? ? inet 192.168.100.2/24 brd 192.168.100.255 scope global br0
[[email protected] ~]# ip r
default via 192.168.32.2 dev ens33
169.254.0.0/16 dev ens33? scope link? metric 1002
192.168.32.0/24 dev ens33? proto kernel? scope link? src 192.168.32.132
192.168.100.0/24 dev br0? proto kernel? scope link? src 192.168.100.2
[[email protected] ~]# ping 192.168.100.2
PING 192.168.100.2 (192.168.100.2) 56(84) bytes of data.
64 bytes from 192.168.100.2: icmp_seq=1 ttl=64 time=0.045 ms
64 bytes from 192.168.100.2: icmp_seq=2 ttl=64 time=0.167 ms
^C
--- 192.168.100.2 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1042ms
rtt min/avg/max/mdev = 0.045/0.106/0.167/0.061 ms
[[email protected] ~]# ping 192.168.100.3
PING 192.168.100.3 (192.168.100.3) 56(84) bytes of data.
From 192.168.100.2 icmp_seq=1 Destination Host Unreachable
From 192.168.100.2 icmp_seq=2 Destination Host Unreachable
From 192.168.100.2 icmp_seq=3 Destination Host Unreachable
^C
--- 192.168.100.3 ping statistics ---
6 packets transmitted, 0 received, +3 errors, 100% packet loss, time 5138ms
pipe 4
[[email protected] ~]#

虚机2

[[email protected] ~]# ip a|grep br0
5: br0: <BROADCAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN qlen 1000
? ? inet 192.168.100.3/24 brd 192.168.100.255 scope global br0
[[email protected] ~]# ip r
default via 192.168.32.2 dev ens33
169.254.0.0/16 dev ens33? scope link? metric 1002
169.254.0.0/16 dev br0? scope link? metric 1005
192.168.32.0/24 dev ens33? proto kernel? scope link? src 192.168.32.134
192.168.100.0/24 dev br0? proto kernel? scope link? src 192.168.100.3
[[email protected] ~]# ping 192.168.100.3
PING 192.168.100.3 (192.168.100.3) 56(84) bytes of data.
64 bytes from 192.168.100.3: icmp_seq=1 ttl=64 time=0.066 ms
64 bytes from 192.168.100.3: icmp_seq=2 ttl=64 time=0.095 ms
^C
--- 192.168.100.3 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1042ms
rtt min/avg/max/mdev = 0.066/0.080/0.095/0.017 ms
[[email protected] ~]# ping 192.168.100.2
PING 192.168.100.2 (192.168.100.2) 56(84) bytes of data.
From 192.168.100.3 icmp_seq=1 Destination Host Unreachable
From 192.168.100.3 icmp_seq=2 Destination Host Unreachable
From 192.168.100.3 icmp_seq=3 Destination Host Unreachable
^C
--- 192.168.100.2 ping statistics ---
5 packets transmitted, 0 received, +3 errors, 100% packet loss, time 4092ms
pipe 4
[[email protected] ~]#

8.
上面只是临时配置br0的地址,重启就消失.
可写入网卡配置文件,永久生效.参考见下:

虚机1

[[email protected] ~]# cd /etc/sysconfig/network-scripts/
[[email protected] network-scripts]# pwd
/etc/sysconfig/network-scripts
[[email protected] network-scripts]# cat ifcfg-br0
DEVICETYPE=ovs
TYPE=OVSBridge
BOOTPROTO=none
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes
IPV4_FAILURE_FATAL=no
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_PEERDNS=yes
IPV6_PEERROUTES=yes
IPV6_FAILURE_FATAL=no
IPV6_ADDR_GEN_MODE=stable-privacy
NAME=br0
DEVICE=br0
ONBOOT=yes
IPADDR=192.168.100.2
NETMASK=255.255.255.0
[[email protected] network-scripts]#

虚机2

[[email protected] ~]# cd /etc/sysconfig/network-scripts/
[[email protected] network-scripts]# cat ifcfg-br0
DEVICETYPE=ovs
TYPE=OVSBridge
BOOTPROTO=none
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes
IPV4_FAILURE_FATAL=no
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_PEERDNS=yes
IPV6_PEERROUTES=yes
IPV6_FAILURE_FATAL=no
IPV6_ADDR_GEN_MODE=stable-privacy
NAME=br0
DEVICE=br0
ONBOOT=yes
IPADDR=192.168.100.3
NETMASK=255.255.255.0
[[email protected] network-scripts]#

9.
创建gre隧道

虚机1
网桥br0添加gre1端口和gre1接口,设置端口类型为gre,配置远端Ip为192.168.32.140,远端ip为另一方的ip.

[[email protected] ~]# ovs-vsctl add-port br0 gre1 -- set interface gre1 type=gre option:remote_ip=192.168.32.140

虚机2
网桥br0添加gre1端口和gre1接口,设置端口类型为gre,配置远端Ip为192.168.32.139,远端ip为另一方的ip.

[[email protected] ~]# ovs-vsctl add-port br0 gre1 -- set interface gre1 type=gre option:remote_ip=192.168.32.139

ovs信息

虚机1

[[email protected] ~]# ovs-vsctl show
67ccf09b-d0d9-4ccb-9f2f-1ac7918e5c46
? ? Bridge "br0"
? ? ? ? Port "gre1"
? ? ? ? ? ? Interface "gre1"
? ? ? ? ? ? ? ? type: gre
? ? ? ? ? ? ? ? options: {remote_ip="192.168.32.140"}
? ? ? ? Port "br0"
? ? ? ? ? ? Interface "br0"
? ? ? ? ? ? ? ? type: internal
? ? ovs_version: "2.0.0"
[[email protected] ~]#

虚机2

[[email protected] ~]# ovs-vsctl show
25cdb171-59ee-4eb4-a8af-cd93dc460926
? ? Bridge "br0"
? ? ? ? Port "gre1"
? ? ? ? ? ? Interface "gre1"
? ? ? ? ? ? ? ? type: gre
? ? ? ? ? ? ? ? options: {remote_ip="192.168.32.139"}
? ? ? ? Port "br0"
? ? ? ? ? ? Interface "br0"
? ? ? ? ? ? ? ? type: internal
? ? ovs_version: "2.0.0"
[[email protected] ~]#

隧道建立后,网络实现互通.

[[email protected] ~]# ip a |grep br0
5: br0: <BROADCAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN qlen 1000
? ? inet 192.168.100.2/24 brd 192.168.100.255 scope global br0
[[email protected] ~]# ip r
default via 192.168.32.2 dev ens33
169.254.0.0/16 dev ens33? scope link? metric 1002
169.254.0.0/16 dev br0? scope link? metric 1005
192.168.32.0/24 dev ens33? proto kernel? scope link? src 192.168.32.139
192.168.100.0/24 dev br0? proto kernel? scope link? src 192.168.100.2
[[email protected] ~]# ping 192.168.100.3
PING 192.168.100.3 (192.168.100.3) 56(84) bytes of data.
64 bytes from 192.168.100.3: icmp_seq=1 ttl=64 time=3.61 ms
64 bytes from 192.168.100.3: icmp_seq=2 ttl=64 time=1.61 ms
64 bytes from 192.168.100.3: icmp_seq=3 ttl=64 time=0.936 ms
^C
--- 192.168.100.3 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 0.936/2.056/3.613/1.135 ms
[[email protected] ~]#
[[email protected] ~]# ip a |grep br0
5: br0: <BROADCAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN qlen 1000
? ? inet 192.168.100.3/24 brd 192.168.100.255 scope global br0
[[email protected] ~]# ip r
default via 192.168.32.2 dev ens33
169.254.0.0/16 dev ens33? scope link? metric 1002
169.254.0.0/16 dev br0? scope link? metric 1005
192.168.32.0/24 dev ens33? proto kernel? scope link? src 192.168.32.140
192.168.100.0/24 dev br0? proto kernel? scope link? src 192.168.100.3
[[email protected] ~]# ping 192.168.100.2
PING 192.168.100.2 (192.168.100.2) 56(84) bytes of data.
64 bytes from 192.168.100.2: icmp_seq=1 ttl=64 time=3.59 ms
64 bytes from 192.168.100.2: icmp_seq=2 ttl=64 time=3.38 ms
64 bytes from 192.168.100.2: icmp_seq=3 ttl=64 time=0.469 ms
^C
--- 192.168.100.2 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2002ms
rtt min/avg/max/mdev = 0.469/2.483/3.596/1.427 ms
[[email protected] ~]#

10.
创建gre通道后系统重启,br0无法获取ip.

暂时没有找到彻底解决这个问题的方法.怀疑是gre本身的问题导致.
临时测试,可以通过把执行命令写入rc.local,解决掉这个问题,命令见下:

[[email protected] ~]#
[[email protected] network-scripts]# cat /etc/rc.d/rc.local |grep -v ^#
touch /var/lock/subsys/local
ifdown br0 &&ifup br0? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ##加这条命令
[[email protected] network-scripts]#

注意必须保证rc.local具有x权限,可执行命令chmod +x /etc/rc.d/rc.local

虚机2做一样的操作.

11.
互相ping,互相抓包.
必须ping才能抓到包.注意了.
如果没有tcpdump命令,安装即可,参考命令:

yum -y install tcpdump

虚机1br0端口抓包icmp

[[email protected] ~]# tcpdump -i br0 icmp
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on br0, link-type EN10MB (Ethernet), capture size 262144 bytes
14:38:48.871368 IP 192.168.100.3 > ovs1: ICMP echo request, id 2235, seq 10, length 64
14:38:48.871442 IP ovs1 > 192.168.100.3: ICMP echo reply, id 2235, seq 10, length 64
14:38:49.873007 IP 192.168.100.3 > ovs1: ICMP echo request, id 2235, seq 11, length 64
14:38:49.873059 IP ovs1 > 192.168.100.3: ICMP echo reply, id 2235, seq 11, length 64
14:38:50.874202 IP 192.168.100.3 > ovs1: ICMP echo request, id 2235, seq 12, length 64
14:38:50.874286 IP ovs1 > 192.168.100.3: ICMP echo reply, id 2235, seq 12, length 64
14:38:51.876210 IP 192.168.100.3 > ovs1: ICMP echo request, id 2235, seq 13, length 64
14:38:51.876272 IP ovs1 > 192.168.100.3: ICMP echo reply, id 2235, seq 13, length 64
14:38:52.878047 IP 192.168.100.3 > ovs1: ICMP echo request, id 2235, seq 14, length 64
14:38:52.878112 IP ovs1 > 192.168.100.3: ICMP echo reply, id 2235, seq 14, length 64
14:38:53.880465 IP 192.168.100.3 > ovs1: ICMP echo request, id 2235, seq 15, length 64
14:38:53.880496 IP ovs1 > 192.168.100.3: ICMP echo reply, id 2235, seq 15, length 64
14:38:54.881610 IP 192.168.100.3 > ovs1: ICMP echo request, id 2235, seq 16, length 64
14:38:54.881681 IP ovs1 > 192.168.100.3: ICMP echo reply, id 2235, seq 16, length 64
^C
14 packets captured
14 packets received by filter
0 packets dropped by kernel

虚机2br0端口抓包icmp

[[email protected] ~]# tcpdump -i br0 icmp
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on br0, link-type EN10MB (Ethernet), capture size 262144 bytes
14:40:13.642693 IP 192.168.100.2 > ovs2: ICMP echo request, id 2253, seq 1, length 64
14:40:13.642725 IP ovs2 > 192.168.100.2: ICMP echo reply, id 2253, seq 1, length 64
14:40:14.642769 IP 192.168.100.2 > ovs2: ICMP echo request, id 2253, seq 2, length 64
14:40:14.642810 IP ovs2 > 192.168.100.2: ICMP echo reply, id 2253, seq 2, length 64
14:40:15.643391 IP 192.168.100.2 > ovs2: ICMP echo request, id 2253, seq 3, length 64
14:40:15.643427 IP ovs2 > 192.168.100.2: ICMP echo reply, id 2253, seq 3, length 64
14:40:16.645941 IP 192.168.100.2 > ovs2: ICMP echo request, id 2253, seq 4, length 64
14:40:16.645992 IP ovs2 > 192.168.100.2: ICMP echo reply, id 2253, seq 4, length 64
14:40:17.647337 IP 192.168.100.2 > ovs2: ICMP echo request, id 2253, seq 5, length 64
14:40:17.647410 IP ovs2 > 192.168.100.2: ICMP echo reply, id 2253, seq 5, length 64
14:40:18.650543 IP 192.168.100.2 > ovs2: ICMP echo request, id 2253, seq 6, length 64
14:40:18.650606 IP ovs2 > 192.168.100.2: ICMP echo reply, id 2253, seq 6, length 64
^C
12 packets captured
12 packets received by filter
0 packets dropped by kernel

虚机1ens33端口抓gre包

[[email protected] ~]#? tcpdump -i ens33 ‘proto gre‘ -n
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on ens33, link-type EN10MB (Ethernet), capture size 262144 bytes
14:46:39.388735 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo request, id 2271, seq 22, length 64
14:46:39.388786 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo reply, id 2271, seq 22, length 64
14:46:40.389953 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo request, id 2271, seq 23, length 64
14:46:40.390074 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo reply, id 2271, seq 23, length 64
14:46:41.390733 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo request, id 2271, seq 24, length 64
14:46:41.390852 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo reply, id 2271, seq 24, length 64
14:46:42.393785 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo request, id 2271, seq 25, length 64
14:46:42.393947 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo reply, id 2271, seq 25, length 64
14:46:43.396383 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo request, id 2271, seq 26, length 64
14:46:43.396545 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo reply, id 2271, seq 26, length 64
^C
10 packets captured
10 packets received by filter
0 packets dropped by kernel

虚机2ens33端口抓gre包

[[email protected] ~]#? tcpdump -i ens33 ‘proto gre‘ -n
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on ens33, link-type EN10MB (Ethernet), capture size 262144 bytes
14:47:57.785331 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo request, id 2298, seq 6, length 64
14:47:57.785451 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo reply, id 2298, seq 6, length 64
14:47:58.789091 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo request, id 2298, seq 7, length 64
14:47:58.789169 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo reply, id 2298, seq 7, length 64
14:47:59.787219 IP 192.168.32.139 > 192.168.32.140: GREv0, length 46: ARP, Request who-has 192.168.100.3 tell 192.168.100.2, length 28
14:47:59.787618 IP 192.168.32.140 > 192.168.32.139: GREv0, length 46: ARP, Reply 192.168.100.3 is-at 12:8f:8c:8f:a8:4e, length 28
14:47:59.789472 IP 192.168.32.139 > 192.168.32.140: GREv0, length 102: IP 192.168.100.2 > 192.168.100.3: ICMP echo request, id 2298, seq 8, length 64
14:47:59.789573 IP 192.168.32.140 > 192.168.32.139: GREv0, length 102: IP 192.168.100.3 > 192.168.100.2: ICMP echo reply, id 2298, seq 8, length 64
^C
8 packets captured
8 packets received by filter
0 packets dropped by kernel

由以上抓包可见,配置实现了gre隧道的连通.

原文地址：https://blog.51cto.com/goome/2418128

时间： 2024-10-30 06:09:56

ovs实践1:基础概念和gre隧道实践操作的相关文章

Linux下的GRE隧道及其路由转发

隧道,字面上来看就是一条通道,这条通道由点到点,独立与其他.linux下的隧道其他的了解不深,单独写下最近搭建过的gre隧道和路由转发功能实现. 先说一下隧道的基本概念: 一种技术(协议)或者策略的两个或多个子网穿过另一种技术(协议)或者策略的网络实现互联,称之为overlay topology,这一技术是电信技术的永恒主题之一.     电信技术在发展,多种网络技术并存,一种技术的网络孤岛可能需要穿过另一种技术的网络实现互联,这种情况如果发生在高层协议的PDU封装于低层协议PDU中时通常称之为

linux基础概念和个人笔记总结（6）

防伪码:曾经梦想仗剑走天涯,回过头,却发现还是放不下家的牵挂 后续理论的验证图文详解,我会更新,希望大家可以参考 第十一章 FTP文件传输服务 1.FTP连接 控制连接:tcp21端口,用于发送ftp命令信息 数据连接:tcp20端口,用于上传.下载数据 数据连接的建立类型: a.主动模式:服务器从20端口主动向客户端发起连接 b.被动模式:服务端在指定范围内某个端口被动等待客户端连接 2.FTP用户的类型 匿名用户:anonymous或ftp 本地用户:账号名称.密码等信息都保存在passwd

LVS 实验笔记1 一些基础概念

基础概念 rsync 同步软件 效率不高 inotify 通知 三种集群: LB: load balance 提高服务器的并发能力 HA: 高可用 high availability不会因为一台服务器宕机导致服务不可用 HP:high perfomence  高性能计算集群 并行处理集群 分布式文件系统 将大任务切割为小任务,分别进行处理 health check:健康检查 NFS:net filesystem 共享存储设备 DAS:直接附加存储 块级别交换数据 直接连到多态主机,性能好,但是要

伯克利教授Stuart Russell：人工智能基础概念与34个误区

伯克利教授Stuart Russell:人工智能基础概念与34个误区 机器之心 9 个月前 机器之心 Russell 是加州大学伯克利分校人工智能系统中心创始人兼计算机科学专业教授,同时还是人工智能领域里「标准教科书」<人工智能:一种现代方法>作者(谷歌研究主管 Peter Norvig 也是该书作者).在这篇文章中,他以 Q&A 的方式讲解了人工智能的未来以及常见的误解. 1. 什么是人工智能? 是对让计算机展现出智慧的方法的研究.计算机在获得正确方向后可以高效工作,在这里,正确的方

【Machine Learn】机器学习及其基础概念简介

机器学习及其基础概念简介 作者:白宁超 2016年12月23日21:24:51 摘要:随着机器学习和深度学习的热潮,各种图书层出不穷.然而多数是基础理论知识介绍,缺乏实现的深入理解.本系列文章是作者结合视频学习和书籍基础的笔记所得.本系列文章将采用理论结合实践方式编写.首先介绍机器学习和深度学习的范畴,然后介绍关于训练集.测试集等介绍.接着分别介绍机器学习常用算法,分别是监督学习之分类(决策树.临近取样.支持向量机.神经网络算法)监督学习之回归(线性回归.非线性回归)非监督学习(K-means聚

《Spark大数据分析：核心概念、技术及实践》大数据技术一览

本节书摘来自华章出版社<Spark大数据分析:核心概念.技术及实践>一书中的第1章,第1节,作者穆罕默德·古勒(Mohammed Guller)更多章节内容可以访问云栖社区"华章计算机"公众号查看. 大数据技术一览 我们正处在大数据时代.数据不仅是任何组织的命脉,而且在指数级增长.今天所产生的数据比过去几年所产生的数据大好几个数量级.挑战在于如何从数据中获取商业价值.这就是大数据相关技术想要解决的问题.因此,大数据已成为过去几年最热门的技术趋势之一.一些非常活跃的开源项目都

Docker 基础概念科普 和 常用操作介绍

Docker 基础概念 Docker是什么? Docker的思想来自于集装箱,集装箱解决了:在一艘大船上,可以把货物规整的摆放起来.并且各种各样的货物被集装箱标准化了,集装箱和集装箱之间不会互相影响.那么我就不需要专门运送瓷器的船和专门运送化学物品的船了.只要这些货物在集装箱里封装的好好的,那我就可以用一艘大船把他们都运走. 不同的应用程序可能会有不同的应用环境,比如.net开发的网站和php开发的网站依赖的软件就不一样,如果把他们依赖的软件都安装在一个服务器上就要调试很久,而且很麻烦,还会造成

2019-04-18-NFV基础概念

NFV技术的起源和概念 在移动互联网时代,运营商面临内外困局.就自身而言,采用的流量增长-网络扩容-收入增长的商业模型正在失效,庞大.僵化的电信基础网络,不能够满足用户的丰富需求:就竞争对手而言,互联网企业以天为计的业务迭代时间,能够很好地贴合用户需求,飞速发展的OTT业务,使运营商越来越趋向于管道工的角色. 随着云计算普及和x86服务器性能提高,各大运营商为了避免进一步成为哑管道的尴尬,由全球各大运营商牵头提出网络功能虚拟化技术(Network Function Virtualization,

[转帖]Zookeeper学习系列【一】 教会你Zookeeper的一些基础概念

Zookeeper学习系列[一] 教会你Zookeeper的一些基础概念 https://segmentfault.com/a/1190000018927058 前言 最近加入了部门的技术兴趣小组,被分配了Zookeeper的研究任务.在研究过程当中,发现Zookeeper由于其开源的特性和其卓越的性能特点,在业界使用广泛,有很多的应用场景,而这些不同的应用场景实际上底层的原理都是差不多的,只要你真正理解了Zookeeper的一些基础概念和机制,就能够触类旁通. 于是乎,在第一次和项目小组内成员