RE: "professional" in an IETF context

[Vasilenko Eduard <vasilenko.eduard@xxxxxxxxxx>]

What about WiFi or 3GPP? Should it increase MTU too?

Reminder: wireless is a loss environment, repeating 64k frames would be very expensive.

 

And if the majority of links (by number) would be still limited, how big is the value for the minority of links supporting hyper-big frames?

 

The problem is not with bandwidth. 16B/800B=2% is not a big overhead.

The problem is with the cost of processing additional bits in ASIC (for Routing, Filtering).

Much more gates are needed on ASICs (including memory chips).

More cost, more power consumption.

 

The situation is especially strange because only 8bytes of IPv6 addresses are used now.

 

Ed/

From: ietf [mailto:ietf-bounces@xxxxxxxx] On Behalf Of Phillip Hallam-Baker
Sent: Thursday, November 4, 2021 6:40 AM
To: Brian E Carpenter <brian.e.carpenter@xxxxxxxxx>
Cc: IETF Discussion Mailing List <ietf@xxxxxxxx>
Subject: Re: "professional" in an IETF context

 

 

 

On Mon, Nov 1, 2021 at 3:49 PM Brian E Carpenter <brian.e.carpenter@xxxxxxxxx> wrote:

Somebody whose email never reaches my inbox alledgedly said:

>       > IPv6 with unnecessarily lengthy 16B addresses without valid
>       > technical reasoning only to make network operations prohibitively
>       > painful is a garbage protocol.

Apart from its incivility, this sentence is factually untrue. The address
length was 8 bytes in the early design of what became IPv6, which was of
course essential to overcome the main limitation of IPv4. It was expanded
to 16 bytes when the value of an interface identifier in addition to
a routeable prefix was considered. That idea was based on existing
practice in several non-IP network technologies, and on the IPng
requirements process. In other words, on technical reasoning and on
running code.

Professionalism includes factual accuracy.

    Brian

 

+1

 

Remember that only 8 bytes of the 'Internet P' address actually route on the Inter-network.

 

The 16 bit port field for IPv4 is also insufficient. But given the choice, I would much rather have 64 bits I can use for local routing than an 80 bit port field.

 

The size of an IPv6 header is only an issue because the Ethernet standard has failed to mandate support for larger frames to keep up with needs. 1500 bytes of payload is too small. Problem is that until a new version of ethernet mandate support for decent sized frames, using them is going to pose practical difficulties.

 

They should increase the frame size to 64KB and limit the scope of the checksum to the first 500 bytes. Payload integrity is not a layer 3 concern. Misrouting due to corrupted address headers is though.