L******t
发帖数: 1985 | 1 On average for a mid/high-end router, how many nexthops (or adjacencies)
does it maintain to its immediate neighbors? A few hundreds? Or a few
thousand?
A BRAS can maintain a lot. But the focus here is on aggregate/core layer
router. A high # can be a major overhead. Just wanted some real case numbers
. Thanks for your input! |
|
L******t
发帖数: 1985 | 2 So realistically speaking, considering all physical interfaces and
subinterfaces, a router usually has no more than few hundreds nexthops (or
somewhat ARP/NDP entries) to maintain, correct? Network engineers just won't
do something beyond that.
year |
|
I********x
发帖数: 858 | 3 Your question is not very clear. For same prefix normally there wont be many
different nexthop. But a router may have many different prefixes and each
prefix can have different next hop. For example, a Internet BGP router
should have more than 600k prefixes, but of course the next hop maybe same
for some of these prefix.
In old days, a ppp dialer router may have thousands of subscribers with
different next hop; now ethernet router can have 60k+ arp entries, depends
on the hardware.
't |
|
L******t
发帖数: 1985 | 4 Not talking about RIB/FIB prefixes. But direct neighbors equivalent to ARP/
NDP entries. Nexthop/Adjacency can mean remote peers. But here it only means
direct ones maintained at Layer 2.
Wanted to know the common practice. Sometimes QA simulate traffic from
thousands of directly connected hosts, which I do not think happens in any
real network.
many |
|
l***y
发帖数: 791 | 5 In the rfc it says mpls forwarding state means either in-lable/outlabel-
nexthop for non-ingress or fec/out-label-nexthop for ingress, but what is '
next hop' in this context? Interface or next hop neighbor ip?
So if IGP updates but for some reason nothing updates ldp, i.e. if ospf
routes around a router with dead RE because ospf timed out but nothing else
has gone down (ldp has much larger hold timer), would mpls forwarding still
happen towards this router with dead RE?
Also, i'm reading about ... 阅读全帖 |
|
|
s*****g
发帖数: 1055 | 7 No need to be per-packet load balancing and it is undesirable. Depends how
the router does the hash, per flow load-balancing would achieve higher
throughput if multitasking -- which is very common.
OP just needs a small router which has three interfaces, one to PC, the
other two will get (DHCP) IP addresses from two ISPs, then configure two
static routes on the router with nexthop pointing to ISP's gateway.
load |
|
m*******r
发帖数: 1 | 8 现有两台C类网段的路由器，路由器1的网段为192.168.10.0，网关为192.168.10.1；路
由器2的网段为192.168.20.0，网关为192.168.20.1，且路由器2的WAN口接在路由器1的
LAN上，并获得ip地址192.168.10.100。从192.168.20.x的机器上访问192.168.10.x的
机器没有问题。为了使192.168.10.x的机器能够访问192.168.20.x的机器，在路由器1
上设置了如下静态路由：
目的IP地址 子网掩码 网关
192.168.20.0 255.255.255.0 192.168.10.100
# ping 192.168.20.1
PING 192.168.20.1 (192.168.20.1) 56(84) bytes of data.
From 192.168.10.1: icmp_seq=1 Redirect Host(New nexthop: 192.168.10.104)
然后就停在这里了，ctrl+c打断，得到下面的输出： |
|
s******v
发帖数: 4495 | 9 youtube被google收购之前，是买level three的service，那时候是$1m/mth左右，这是
根据youtube的拿到的funding，刨去工资之类开销，大概算的。
google有一个自己的全球的网，是不是在很多的地方，和很多的providers有peering？
为什么这个arbor有这个数据啊？bgp updates? netflow? 应该是bgp，根据nexthop?
as path推算出，有% traffic是在where deliver到google network上？
的带宽成本几乎为零，甚至比一个普通个人用户还低。
最新报告显示，YouTube的带宽成本几乎为零。
，认为YouTube的带宽成本约为3.6亿美元，今年有可能亏损5亿美元。但是Arbor
Networks却认为，YouTube的带宽成本甚至比一个普通个人用户还低。虽然在YouTube的
帮助下，谷歌目前在整个互联网流量中的占比已经至少达到6%，而且会使用大量的光缆
，但是谷歌却可以与大型ISP(互联网服务提供商)进行流量交换，而无须支付带宽费用。
安全局以外，对互联网流量最为了解的一 |
|
s******v
发帖数: 4495 | 10 netflix当然没有网络，这个可能是类似bgp peering，只是传输prefix + nexthop |
|
m**k
发帖数: 290 | 11 $ traceroute 69.53.236.17
...
8 tge8-4.fr4.sjc.llnw.net (69.28.171.118) 39.253 ms 22.797 ms tge18-2.
fr3.sjc.llnw.net (69.28.172.54) 25.082 ms
9 netflix.tge7-4.fr4.sjc.llnw.net (69.28.179.74) 48.394 ms ve5.fr4.sjc.
llnw.net (69.28.171.210) 19.060 ms 29.856 ms
10 netflix.tge7-4.fr4.sjc.llnw.net (69.28.179.74) 31.160 ms 25.653 ms 24
.756 ms
11 te1-8.csrt-agg01.prod1.netflix.com (69.53.225.6) 22.401 ms 25.522 ms
xe-2-2-0-954.jnrt-edge01.prod1.netflix.com (69.53.225.26) 32.865 ms
12 ... 阅读全帖 |
|
I********x
发帖数: 858 | 12 I have never used more than four next hops for one prefix during my 15 year
networking experience.
numbers |
|
I********x
发帖数: 858 | 13 On one of my living router/switch, I have 11.5k arp entries. I have about 12
of these routers.
In my case these hosts are connected via tens of layer2 switches but routing
is done on this device.
means |
|
