Anda di halaman 1dari 19

IEEE CCW 08

New Network Architectures: Why Bother?

Paul Francis Cornell

What is a home run in systems research? Intellectually compelling idea

Huge impact on industry


Something that industry wouldnt have done on its own

A home run

Intellectual

Impact

Not done by Industry

As industry matures, What is a home these run in systems get increasingly research? difficult

Intellectually compelling idea

Huge impact on industry


Something that industry wouldnt have done on its own

This is new architecture (clean slate) research

Intellectual

Impact

Not done by Industry

Is it possible to hit a home run in networking research these days? Maybe. But only with incrementally deployable ideas. Dirty slate research....
Finding a fun project that industry wants to deploy is itself an intellectual challenge....

The challenge of impact in network research:


Need buy-in from providers, vendors, and standards

One approach is to become a vendor


Do a startup

But this is not always possible


Inter-domain routing, for instance

One Bottleneck at a Time


Our approach to inter-domain routing research

Dont solve every problem at once


Rather, solve the current most serious problem, move on

Virtual Aggregation (ViAggre)


Shrinks the BGP FIB (by easily 10x), but leaves the RIB intact
ISPs sometimes have to replace hardware because of FIB growth

Intact RIB means no real change to how routing is done


In fact, ViAggre requires no changes to router software!

Today: All router FIBs have routes to all destinations

Dest Next Hop 20.5/16 1.1.1.1 36.3/16 2.1.1.1 ....

Virtual Aggregation: FIBs have routes to only part of the address space
Virtual Prefixes

Dest Next Hop 20.5/16 1.1.1.1 ....


Dest Next Hop 188.3/16 2.1.1.1 ....

Paths through the ISP have two components:


1: Route to a nearby Aggregation Point 2: Tunnel to the neighbor router

1: Routing to a nearby Aggregation Point


Configure Aggregation Point with static route for the Virtual Prefix Virtual Prefix is advertised into BGP

2: Tunnel packet to neighbor router (MPLS)


Static routes for all neighbors are imported into OSPF

MPLS LDP creates tunnels to every neighbor router

We thought we could bypass vendors and standards


Providers could deploy this on their own

Turns out, providers are nervous about doing anything without vendor blessing Fortunately, a vendor (Huawei) became interested in this

Standardizing ViAggre in IETF (IDR)


Going well, because no changes to BGP

Huawei is implementing it With RFC in hand, can try to get providers to convince other vendors to implement

Assuming FIB is solved, whats the next bottleneck? I suspect that it is not RIB size, but rather BGP update processing cost
We are starting some router measurements to find out

Can we reduce the cost of updates while running BGP more-or-less as is?

Mapped-BGP
Expense of route processing are all the policies
Filtering, best-path selection, loadbalance, aggregation, route policies

Our goal is to get rid of the policies for most prefixes

Rather than distribute routes to all prefixes, distribute routes to tunnel endpoints, and distribute maps that bind prefixes to tunnel endpoints
Make the maps policy-free Exploit tunnels to improve inter-AS load balance and increase aggregation opportunities

Anda mungkin juga menyukai