[j-nsp] Layer 2 Fragmentation

[Keegan Holley]

There's no such thing as Ethernet fragmentation.  There's no sequence number field or next frame packet headers and no control message protocol to control it.  You could create a gre tunnel and set the MTU and/or fragmentation there.

Sent from my iPhone

On Mar 6, 2011, at 6:41 AM, Mohammad Salbad <salbad1981 at hotmail.com> wrote:

> Hello All
> 
> 
> 
> I have two EX4200 switches connected back to back to each other with
> interfaces in trunk mode as following:
> 
> Case(1) back to back connection using ethernet cable
> 
> (VL11 IP: 10.1.1.1/30) EX4200 -------------------TRUNK
> (Ethernet)---------------------EX4200 (VL11 IP: 10.1.1.2/30)
> 
> è For this case I can ping with don’t fragment option up to 1554 byte as
> following:
> 
> ping 10.1.1.2 size 1554 do-not-fragment    
> 
> PING 10.1.1.2 (10.1.1.2): 1554 data bytes
> 
> 1562 bytes from 10.1.1.2: icmp_seq=0 ttl=64 time=3.639 ms
> 
> 1562 bytes from 10.1.1.2: icmp_seq=1 ttl=64 time=3.550 ms
> 
> 1562 bytes from 10.1.1.2: icmp_seq=2 ttl=64 time=3.556 ms
> 
> Case(2) back to back connection using wireless backhaul devices
> 
> (VL11 IP: 10.1.1.1/30) EX4200 -------------------TRUNK (wireless
> backhaul)---------------------EX4200 (VL11 IP: 10.1.1.2/30)
> 
> è For this case I can ping with don’t fragment option up to 1476 byte as
> following:
> 
> ping 10.1.1.2 size 1476 do-not-fragment    
> 
> PING 10.1.1.2 (10.1.1.2): 1476 data bytes
> 
> 1484 bytes from 10.1.1.2: icmp_seq=0 ttl=64 time=4.096 ms
> 
> 1484 bytes from 10.1.1.2: icmp_seq=1 ttl=64 time=4.094 ms
> 
> 1484 bytes from 10.1.1.2: icmp_seq=2 ttl=64 time=5.496 ms
> 
> 
> 
> Switches configuration sample
> 
> ge-0/0/1 {
> 
>    mtu 1600;
> 
>    ether-options {
> 
>        link-mode full-duplex;
> 
>        speed {
> 
>            100m;
> 
>        }
> 
>    }
> 
>    unit 0 {
> 
>        family ethernet-switching {
> 
>            port-mode trunk;
> 
>            vlan {
> 
>               members [ TEST ];
> 
>        }
> 
>    }
> 
> }
> 
> vlan {
> 
>    mtu 1600;
> 
>    unit 11 {
> 
>        family inet {
> 
>            address 10.1.1.1/30;
> 
>    }
> 
> }
> 
> vlans {
> 
>    TEST {
> 
>        vlan-id 11;
> 
>    }
> 
> 
> 
> The case is that I didn’t change the configuration on the switches I just
> replaced the Ethernet cable with wireless backhaul devices and then I can’t
> ping with larger mtu sizes, and this is because the wireless devices I’m
> using for backhaul are limited to 1500 MTU size, so I’m wondering if I there
> is anything I can do on the switches to overcome the limitation on the
> wireless backhaul devices, i.e. if I can fragment the packet into smaller
> pieces before sending them to the wireless device and combining the packet
> on the other end (something like link fragmentation and interleaving).
> 
> 
> 
> Thank you
> 
> Mohammad Salbad
> 
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp