1 |
* Grant <emailgrant@×××××.com> [160920 08:53]: |
2 |
> >> > A while back I was having networking issues. I eventually tried |
3 |
> >> > drastically lowering the MTU of all the systems onsite and the issues |
4 |
> >> > disappeared. I always thought the issue was due to the MTU on our |
5 |
> >> > modem/router. Today I read that AT&T DSL requires a 1492 MTU so I |
6 |
> >> > increased the MTU of our systems up to 1492 and haven't had any |
7 |
> >> > issues. Do certain ISPs require you to change the MTU of your entire |
8 |
> >> > network, or is this likely due to our AT&T modem/router itself? |
9 |
> >> |
10 |
> >> AFAIK the MTU is defined for every network interface separately. For an |
11 |
> >> ADSL connection it is common that a lower MTU is needed because of the |
12 |
> >> PPPoE header information that is encapsulated in the ethernet frames. |
13 |
> >> But in that case it is sufficient to lower the MTU just for the WAN |
14 |
> >> interface that is connected to the DSL modem. |
15 |
> >> If you don't use protocol encapsulation in your LAN then there should |
16 |
> >> be IMHO no reason for lowering the MTU of your internal interfaces. |
17 |
> >> |
18 |
> >> -- |
19 |
> >> Regards |
20 |
> >> wabe |
21 |
> > |
22 |
> > MTU is per network interface but you really don't want to end up having |
23 |
> > your router fragment every IP packet because systems on your subnet are |
24 |
> > using a larger MTU. |
25 |
> > |
26 |
> > Todd |
27 |
> |
28 |
> |
29 |
> That makes sense. So in my case, I'm thinking 1492 MTU on every |
30 |
> interface in the network. |
31 |
> |
32 |
> So I'm sure I understand, should everyone with a DSL connection set an |
33 |
> MTU of 1492 (or potentially lower) on all of their network interfaces |
34 |
> to avoid packet fragmentation? |
35 |
> |
36 |
> - Grant |
37 |
|
38 |
I would probably set the MTU to 1492 on each interface myself. |
39 |
|
40 |
But it really depends upon the traffic mix and how "smart" ("dumb") the |
41 |
devices on the network are. |
42 |
|
43 |
TCP is likely using Path MTU Discovery to determine the TCP Maximum |
44 |
Segment Size so that TCP traffic doesn't encounter IP fragmentation |
45 |
end-to-end. PMTU used to use ICMP packets to determine the end-to-end |
46 |
MTU, but RFC4281 (I think) uses a method to work around the dropping or |
47 |
filtering of ICMP packets. |
48 |
|
49 |
However, there's different quality of implementations as well as |
50 |
differences in whether the network stack uses the PMTUD for UDP and |
51 |
other datagram traffic or not as well. |
52 |
|
53 |
Todd |
54 |
|
55 |
DISCLAIMER: It's been a number of years since I've been involved in |
56 |
implementing any IP networking so things may have moved on since then. |