Re: Minimum length of DHCP packet ?

Eric Weber <weber@tungsten.seattle.geoworks.com> Fri, 08 March 1996 17:03 UTC

Received: from ietf.cnri.reston.va.us by IETF.CNRI.Reston.VA.US id aa14908; 8 Mar 96 12:03 EST
Received: from CNRI.Reston.VA.US by IETF.CNRI.Reston.VA.US id aa14903; 8 Mar 96 12:03 EST
Received: from coral.bucknell.edu by CNRI.Reston.VA.US id aa09660; 8 Mar 96 12:03 EST
Received: from charcoal-gw.eg.bucknell.edu by coral.bucknell.edu; (5.65v3.0/1.1.8.2/29Aug94-0956AM) id AA18780; Fri, 8 Mar 1996 11:54:47 -0500
Received: from reef.bucknell.edu by charcoal (5.x/SMI-SVR4) id AA04808; Fri, 8 Mar 1996 11:48:03 -0500
Received: from fusion.geoworks.com by reef.bucknell.edu with SMTP (5.65/IDA-1.2.8) id AA14226; Fri, 8 Mar 1996 11:47:53 -0500
Received: from tungsten.seattle.geoworks.com by geoworks.com (4.1/SMI-4.1) id AA05668; Fri, 8 Mar 96 08:46:30 PST
Received: from astatine.seattle.geoworks.com.seattle by tungsten.seattle.geoworks.com (4.1/SMI-4.0) id AA07395; Fri, 8 Mar 96 08:41:25 PST
Date: Fri, 8 Mar 96 08:41:25 PST
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Eric Weber <weber@tungsten.seattle.geoworks.com>
Message-Id: <9603081641.AA07395@tungsten.seattle.geoworks.com>
To: Mike Carney - Sun BOS Software <Mike.Carney@east.sun.com>
Cc: tomy@sfc.wide.ad.jp, dhcp-v4@bucknell.edu
Subject: Re: Minimum length of DHCP packet ?
In-Reply-To: <9603081626.AA05382@poori.East.Sun.COM>
References: <9603081626.AA05382@poori.East.Sun.COM>

Akihiro Tominaga writes:
> DHCP message should be greater than 300 octets (= BOOTP Packet), is it
> correct?  Or is it harmful to discard packets shorter than 300 octets?

Mike Carney writes:
 > Since DHCP is intended as "BOOTPv2", and thus leverage the already BOOTP 
 > relay agent capabilities out there, the size should be greater than or 
 > equal to 300 bytes. I've seen some relay agents discard packets less than 
 > 300 bytes (the BOOTP spec specifies that BOOTP packets must be 300 bytes in 
 > length).

I see no reason why anybody should deliberately disregard a packet
smaller than 300 octets, if it is large enough to contain a valid DHCP
message.  While it certainly makes sense for clients to pad everything
they send out to 300 octets, I don't see why it makes sense for
anybody to reject an incoming message because it came from a client
that didn't do such padding.

Therefore I would advise the original poster NOT to discard packets
shorter than 300 octets, if they can otherwise be parsed.

-- Eric Weber
   weber@geoworks.com