Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: [EFM] OAM developing Geoff's observation.



Title: RE: [EFM] OAM developing Geoff's observation.

Fletcher,
respected members of the committee,

>> Note, there is not an option between IP and PPP.  IP is a layer 3 protocol,
>> PPP is a data link protocol for IP that we are replacing with Ethernet.

Why replace PPP? The strength of any new technology is in offering new unique advantages while still integrating with existing technologies. EFM will be offering new features that current 802.3 based solutions do not offer, but it seems to me that "replacing PPP with Ethernet" or "replacing PPP with EFM" is something that goes a little bit beyond our wildest dreams.

>        Personally, I think PPPoE is a bad design.  It makes me
>slightly ill to think of the people out there who run PPPoE over
>Ethernet over ATM.  Yes, there are ADSL SPs who do this.

Yes they do. And there is a reason for that.
PPP offers something other layer 2 protocols do not.
SPs use existing technologies and equipment to design their networks.
Existing subscriber aggregation equipment implements PPP, L2TP and other technologies to make it possible for SPs to provide variety of services over a variety of layer 1 available to them and still have a unified approach in subscriber termination.

What I mean is . . . unified QoS, security, billing, VPN/VPRN, etc. over xDSL, ISDN, PSTN, LL, etc.
PPPoEoA allows (just one example) to have two completely different subscribers on the same xDSL subscriber line. By "completely different" i mean having different levels of service, belonging to different VP(r)N(s), etc. In a situation when one xDSL line has only one subscriber on it, PPPoA is used at the moment.

Do you think it is a good idea to replace all that in EFM? That means developing all that functionality in the new protocol? to replace all the existing approaches? To develop new subscriber aggregation equipment for the SPs specifically to address EFM???

Typical SP still has all the PSTN, ISDN, xDSL and LL subscribers . . . So even if EFM is going to offer it's unique set of protocols, technologies and approaches for the subscriber termination, SPs will have to use "old" PPP based approaches for their other (non EFM) subscribers. What makes you think they'll be willing to have two different subscriber aggregation networks instead of one (PPP based)?

I think EFM has to have it's own OAM and some other unique things (basically we need to make 802.3 solutions capable of being used as a subscriber edge), but we should not think about changing all subscriber aggregation technologies here. At least not at this time.

Regards,
Vladimir Senkov.