Hitesh Ballani Paul Francis Tuan Cao and Jia Wang Cornell University and ATampT LabsResearch Presented by Gregory Peaker Zhen Qin Outline Motivation ViAggre design Allocating aggregation points ID: 597990
Download Presentation The PPT/PDF document "Making Routers Last Longer with ViAggre" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Making Routers Last Longer with ViAggre
Hitesh Ballani, Paul Francis, Tuan Cao and Jia Wang
Cornell University and AT&T Labs-Research
Presented by Gregory Peaker, Zhen QinSlide2
OutlineMotivation
ViAggre design
Allocating aggregation points
Evaluation
Deployment
DiscussionSlide3
Motivation
Large Routing Table
More FIB space on Routers
Rapid future growth
IPv4 exhaustion IPv6 deploymentSlide4
Does FIB Size Matter?
Technical
concerns
Power and Heat dissipation problems
Business concerns Large routing table
Less cost-effective networks
Price per bit forwarded increases
Cost of router memory upgrades
ISPs are willing to undergo some pain to extend the life of their routersSlide5
Virtual Aggregation (ViAggre)Slide6
ViAggre basic ideaSlide7
ViAggre basic ideaSlide8
ViAggre basic ideaSlide9
ViAggre basic ideaSlide10
ViAggre basic ideaSlide11
Data-Plane pathsSlide12
Data-Plane pathsSlide13
Ingress Aggregation PointSlide14
Ingress Aggregation PointSlide15
Ingress Aggregation PointSlide16
Aggregation Point EgressSlide17
Aggregation Point EgressSlide18
Aggregation Point EgressSlide19
Aggregation Point EgressSlide20
Allocating aggregation points
A router’s FIB size (F
r
):
routes to the real prefixes in the virtual prefixes it is aggregating
routes to all the virtual prefixesroutes to the popular prefixes
LSP mappings for external routersSlide21
Allocating aggregation points
Traffic stretch:
packets from router
i
to prefix
p
belonging to a virtual prefix
v
are routed through router
k
j
is the egress-router for a traffic from router
k
to prefix
p
i
chooses
k
as an aggregation point that is closest in terms of IGP metrics, where
k
is also belonging to virtual prefix
vSlide22
Allocating aggregation points
Definition of
can_server
If router
i
were to aggregate virtual prefix
v
, which routers can it serve without violating the stretch constraint C.
In accordance with
can_server
relation while trying to minimize the worst FIB size, an algorithm was proposed to designate all routers are served for a virtual prefixSlide23
EvaluationImpact on Traffic
Traffic stretched using different router level path than native path
Increase Router LoadSlide24
Evaluation using ISPsTier 1
Extend life of routers from 2007 to 2018
39% increase load on routers
1.5% of prefixes for 75.5% traffic
5% of prefixes for 90.2% trafficSlide25
Evaluation using ISPsTier 2
Apply routing table for their customers
Use default table for all other customers
Negligible traffic stretch (<0.2 msec)
Negligible Increase in Load (<1.5%)Slide26
DeploymentCan be incrementally deployed
Can be deployed on small scale
Incentive for deployment
No Change to ISP’s routing table
Does not affect routers advertised to neighbors
Does not restrict routing policiesExtra configurationCould be automated
Vendor support + cheaper routersSlide27
DeploymentSlide28
Conclusion & OffenseCan be used by ISPs today
10x reduction in FIB size
Negligible traffic stretch
Negligible load increase
ISPs extend lifetime of routers
A simple and effective first stepSlide29
Discussion & Offense