RE: [RPRWG] RE: payload length and padding and stuff
Tony Jeffree <[email protected]> Thu, 05 Dec 2002 09:45:56 +0000
| Newsgroups | gmane.ietf.iporpr |
|---|---|
| Message-ID | <[email protected]> |
--=====================_162998108==_.ALT Content-Type: text/plain; charset="us-ascii"; format=flowed Dan - Some further clarification here: This very issue came up as a formal Interpretation request on 802.1D and 802.3 recently. The 802.1 agreed interpretation is documented here: http://www.ieee802.org/1/pages/int-2.html Hence, it is the frame size defined by the 802.3 standard (or more generally, the size defined in the standard for the destination port's MAC) that is used by a conformant Bridge in order to identify (and discard) oversize frames. Regards, Tony ------------------------------------------------- Dan - I think that you will find the 802.1D standard requires conformant Bridges to discard frames that exceed the max frame size for the destination LAN: "7.7.1 Enforcing topology restriction Each Port is selected as a potential transmission Port if, and only if a) The Port on which the frame was received was in a forwarding state (8.4), and b) The Port considered for transmission is in a forwarding state, and c) The Port considered for transmission is not the same as the Port on which the frame was received, and d) The size of the mac_service_data_unit conveyed by the frame does not exceed the maximum size of mac_service_data_unit supported by the LAN to which the Port considered for transmission is attached. For each Port not selected as a potential transmission Port, the frame shall be discarded." Regards, Tony At 18:25 04/12/2002 +0200, Romascanu, Dan (Dan) wrote: >Nader, > >There are actually two different layers here. > >802.3 MACs and PHYs MUST support a max frame length of 1522. You may find >implementations that are both standard compliant and support jumbo sizes >as proprietary extensions. > >I think that 802.1 bridges do not have such a limitation within the >standard. As you mention, they can be configured to support larger frame >sizes. Actually some of the non-Ethernet 802 technologies already support >larger frame sizes. > >Dan > > > > -----Original Message----- > > From: Nader Vijeh [mailto:[email protected]] > > Sent: Wednesday, December 04, 2002 6:10 PM > > To: Romascanu, Dan (Dan) > > Cc: [email protected]; [email protected] > > Subject: RE: [RPRWG] RE: [IPORPR] payload length and padding and stuff > > > > > > Dan, > > > > Thank you for the correction. Jumbo frames are not supported > > by the 802.3 > > standard. > > There are Ethernet switches on the market that support jumbo > > frames in a > > proprietary manner. Of course these will not interoperate > > with standard > > compliant 802.1 bridges. As I pointed out in the rest of the > > message, even > > if jumbo frames are supported, max frame size "needs to be > > configurable to > > be lower, in order to comply with transparent bridging requirements". > > > > Nader > > > > -----Original Message----- > > From: Romascanu, Dan (Dan) [mailto:[email protected]] > > Sent: Wednesday, December 04, 2002 12:45 AM > > To: Nader Vijeh; Necdet Uzun; Anoop Ghanwani > > Cc: Frank Kastenholz; [email protected]; [email protected] > > Subject: RE: [RPRWG] RE: [IPORPR] payload length and padding and stuff > > > > > > > > > > The max frame size may be extended beyond 802.3 1522 bytes > > > limit as there is > > > precedence in 802.3 community. > > > > Nader, > > > > I am not sure what you exactly mean. If you refer to what is > > popularly known > > as 'Jumbo frames', they are not supported by the IEEE 802.3 > > standards. > > > > Dan > > Regards, Tony --=====================_162998108==_.ALT Content-Type: text/html; charset="us-ascii" <html> Dan -<br><br> Some further clarification here:<br><br> This very issue came up as a formal Interpretation request on 802.1D and 802.3 recently. The 802.1 agreed interpretation is documented here:<br><br> <a href="http://www.ieee802.org/1/pages/int-2.html" eudora="autourl">http://www.ieee802.org/1/pages/int-2.html</a><br><br> Hence, it is the frame size defined by the 802.3 standard (or more generally, the size defined in the standard for the destination port's MAC) that is used by a conformant Bridge in order to identify (and discard) oversize frames.<br><br> Regards,<br> Tony<br><br> -------------------------------------------------<br> Dan -<br><br> I think that you will find the 802.1D standard requires conformant Bridges to discard frames that exceed the max frame size for the destination LAN:<br><br> "<font face="Arial, Helvetica"><b>7.7.1 Enforcing topology restriction<br> </b></font><font face="Times New Roman, Times">Each Port is selected as a potential transmission Port if, and only if<br> a) The Port on which the frame was received was in a forwarding state (8.4), and<br> b) The Port considered for transmission is in a forwarding state, and<br> c) The Port considered for transmission is not the same as the Port on which the frame was received,<br> and<br> d) The size of the mac_service_data_unit conveyed by the frame does not exceed the maximum size of<br> mac_service_data_unit supported by the LAN to which the Port considered for transmission is<br> attached.<br> For each Port not selected as a potential transmission Port, the frame shall be discarded."<br><br> </font>Regards,<br> Tony<br><br> <br><br> At 18:25 04/12/2002 +0200, Romascanu, Dan (Dan) wrote:<br><br> <blockquote type=cite class=cite cite>Nader,<br><br> There are actually two different layers here. <br><br> 802.3 MACs and PHYs MUST support a max frame length of 1522. You may find implementations that are both standard compliant and support jumbo sizes as proprietary extensions.<br><br> I think that 802.1 bridges do not have such a limitation within the standard. As you mention, they can be configured to support larger frame sizes. Actually some of the non-Ethernet 802 technologies already support larger frame sizes.<br><br> Dan<br><br> <br> > -----Original Message-----<br> > From: Nader Vijeh [<a href="mailto:[email protected]" eudora="autourl">mailto:[email protected]</a>]<br> > Sent: Wednesday, December 04, 2002 6:10 PM<br> > To: Romascanu, Dan (Dan)<br> > Cc: [email protected]; [email protected]<br> > Subject: RE: [RPRWG] RE: [IPORPR] payload length and padding and stuff<br> > <br> > <br> > Dan,<br> > <br> > Thank you for the correction. Jumbo frames are not supported <br> > by the 802.3<br> > standard.<br> > There are Ethernet switches on the market that support jumbo <br> > frames in a<br> > proprietary manner. Of course these will not interoperate <br> > with standard<br> > compliant 802.1 bridges. As I pointed out in the rest of the <br> > message, even<br> > if jumbo frames are supported, max frame size "needs to be <br> > configurable to<br> > be lower, in order to comply with transparent bridging requirements".<br> > <br> > Nader<br> > <br> > -----Original Message-----<br> > From: Romascanu, Dan (Dan) [<a href="mailto:[email protected]" eudora="autourl">mailto:[email protected]</a>]<br> > Sent: Wednesday, December 04, 2002 12:45 AM<br> > To: Nader Vijeh; Necdet Uzun; Anoop Ghanwani<br> > Cc: Frank Kastenholz; [email protected]; [email protected]<br> > Subject: RE: [RPRWG] RE: [IPORPR] payload length and padding and stuff<br> > <br> > <br> > > <br> > > The max frame size may be extended beyond 802.3 1522 bytes <br> > > limit as there is<br> > > precedence in 802.3 community.<br> > <br> > Nader,<br> > <br> > I am not sure what you exactly mean. If you refer to what is <br> > popularly known<br> > as 'Jumbo frames', they are not supported by the IEEE 802.3 <br> > standards. <br> > <br> > Dan<br> > </blockquote> <x-sigsep><p></x-sigsep> Regards,<br> Tony<br> </html> --=====================_162998108==_.ALT--