Susan,

In chapter 2 Definitions there is a copy/paste error I presume


   PWE3: Pseudo Wire Emulation Edge-to-Edge

   EPC: Pseudo Wire Emulation Edge-to-Edge

Presume that EPC = Evolved Packet Core

Have several comments for the rest of the draft, as it is somewhat  not clearly 
written. I can't figure out what is the plan for the access and what for the 
aggregation networks in mobile backhaul.

In Application configuration chapter 3.1,

the draft talks about protecting legacy radio networks, but it doesn't discuss 
that todays MBH are designs, access (hub and spoke) and aggregation networks 
(rings with IP/MPLS and IGP, mostly OSPF, overlay).

In same chapter,

The radio architecture evolution will bring out new radio interfaces,
such as the X2 interface in LTE which will not work in hub-spoke
communication mode and needs much more shorter latency.

Introduction of X2 shortens the latency for control and data traffic, as now 
eNodeBs can communicate directly, not as before via RNC. With X2 interfaces, 
can create mesh topology in MBH access part, which allows multi path to EPC 
from each eNodeB.
With introduction of 5G networks, as spectrum allocation will be dynamic, the 
bandwidth requirements will fluctuate from different areas, depending on the 
allocated spectrum to the eNodeBs.

2. various radio applications
next gen networks (actually this was definition and requirement for 4G) are IP 
networks, so real-time applications like voice and video should be IP 
applications, not really radio applications. The classifications for traffic 
should be real-time, near-realtime non-real time user traffic and control 
traffic. As radio should carry only data and signaling, there should be no 
other differentiation.

3.  various network architectures:

The mobile backhaul network maybe consist of hundreds of nodes
in a small county or thousands of nodes in a populous region.
It will be an integration of different ASNs rather than a
single AS, when EPC is deployed in the Core network with LTE.
The network devices on different points of the network (e.g.
access\aggregation\core) have different routing and protocol
processing capabilities, resulting in an integration of
different IGP routing areas rather than a single large IGP
routing area.  Within various network architectures, different
service modes should be provided, such as SS- PW or MS-PW, E2E
L3VPN or HVPN, Seamless MPLS, and the integration of them.

Can you please clarify this chapter? MBH topologies today are very much 
geographically created and connected to the core, so for example network 
operator in Boston will have access and aggregation network connected to the 
core in Boston and this will be within single AS. The edge device between 
aggregation and core  is connected to multiple ASNs, but not devices with 
access and aggregation networks? Do you expect to see devices in access 
networks to be connected to different ASNs?

In 3.2 are you looking to use BGP as IGP and EGP protocol? Why not run IGP in 
access networks, as those are mesh topologies with X2 interfaces, especially as 
X2 has same S1 characteristics for E-UTRAN service continuation. Also eNodeBs 
will be often connected via wireless connections (microwave), so using an IGP 
in access part of the mobile backhaul should be considered and how to 
manipulate certain metrics based on the wireless backhaul portion performance.

Do you presume running pseudo wires all the way through the access portion of 
MBH or just to the pre-aggregation site?

If you could clarify the network topology in mind, that would be helpful.

I believe there are different requirements for access and aggregation part of 
the mobile backhaul for i2rs and that your draft is more applicable to the 
aggregation part of mobile backhaul.

Regards,

Dean

On Mar 6, 2014, at 5:01 AM, Susan Hares 
<[email protected]<mailto:[email protected]>> wrote:


During the last two weeks, people have indicated interested in mobile backhaul 
cases.    We would appreciate input on the following draft:

https://datatracker.ietf.org/doc/draft-zhang-i2rs-mbb-usecases/

If you are interested in chatting about this draft would you please contact me 
([email protected]<mailto:[email protected]>). For those who are at the IETF and 
would like to talk about MBB over beer, join us at the IETF registration desk 
at 7:30pm (Thursday 3/6/14).

Sue Hares


<Untitled attachment 00301.txt>_______________________________________________
i2rs mailing list
[email protected]<mailto:[email protected]>
https://www.ietf.org/mailman/listinfo/i2rs

_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs

Reply via email to