OptFlow: A Flow-based Abstraction for Programmable Topologies - PowerPoint PPT Presentation

OptFlow: A Flow-based Abstraction for Programmable Topologies Klaus-Tycho Foerster Long Luo Manya Ghobadi Computer Science and Artificial Faculty of Computer Science University of Electronic Science University of Vienna and Technology of China Intelligence Laboratory, MIT Austria P.R. China USA Paper available at https://dl.acm.org/doi/10.1145/3373360.3380840 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 1

Wide/Regional/Metro-Area Networks are not Static OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 2

Wide/Regional/Metro-Area Networks are not Static OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 2

Closer look: Optical components • Dense Wavelength Division Multiplexing (DWDM) - > 100 wavelengths (e.g. 100Gbps ) per fiber edges OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 3

Closer look: Optical components • Dense Wavelength Division Multiplexing (DWDM) - > 100 wavelengths (e.g. 100Gbps ) per fiber edges • Wavelengths can be steered at connection points ◦ By ROADM s (reconfigurable add/drop multiplexers) nodes OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 3

Intuition: Move wavelengths u x v w wavelength OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 4

Intuition: Move wavelengths u x v w wavelength capacity = 1 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 4

Intuition: Move wavelengths u x 1 1 1 1 v w wavelength capacity = 1 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 4

Intuition: Move wavelengths u x u x 1 1 1 1 1 1 1 1 v w v w wavelength capacity = 1 throughput: 1 + 1 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 4

Intuition: Move wavelengths 2 u x u x u x 1 1 1 1 1 1 0 0 2 1 1 v w v w v w wavelength capacity = 1 throughput: 1 + 1 throughput: 2 + 2 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 4

Intuition: Move wavelengths 2 u x u x u x 1 1 1 1 1 1 0 0 2 1 1 v w v w v w wavelength capacity = 1 throughput: 1 + 1 throughput: 2 + 2 demand-aware capacity (wavelengths) OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 4

Intuition: Move wavelengths Topology Programmability (TP) + Traffic Engineering (TE) > TE 2 u x u x u x 1 1 1 1 1 1 0 0 2 1 1 v w v w v w wavelength capacity = 1 throughput: 1 + 1 throughput: 2 + 2 demand-aware capacity (wavelengths) OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 4

How to leverage Topology Programmability (TP)? • Variant #A: Optimize TP and Traffic Engineering (TE) separately? ◦ Inefficient, misses opportunities (recall last slide) • Variant #B: Redesign TE to include TP? ◦ Tedious, operators are reluctant • Variant #C: Don’t change TE, still incorporate TP! ◦ Abstractions! OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 5

How to incorporate current Traffic Engineering (TE)? Abstractions Unmodified Traffic Eng. Flow routing Augmented Topology Reconfiguration OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 6

How to design the Abstraction? u x v w Setting in our example: • Every node supports 2 wavelengths • Every edge supports 2 wavelengths OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 7

How to design the Abstraction? u x u x v w v Setting in our example: Idea: u should only send 2 real units of traffic • • Every node supports 2 wavelengths Implement fake flows that block capacity • • Every edge supports 2 wavelengths Represent dual wavelength assignment • TEs can deal with flows OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 7

How to design the Abstraction? u x u x u x v w v v Setting in our example: Idea: u should only send 2 real units of traffic TE performs TP by routing both flow types • • • Every node supports 2 wavelengths Implement fake flows that block capacity Fake flows from u to x or v • • • Every edge supports 2 wavelengths Represent dual wavelength assignment Real traffic from u to x • TEs can deal with flows OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 7

Intuition for the Abstraction u x u→x : 2 v→w : 2 v w Setting in our example: • Every node supports 2 wavelengths • Every edge supports 2 wavelengths • u wants to send traffic to x • v wants to send traffic to w OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 8

Intuition for the Abstraction u x u x u→x : 2 u→x : 2 u→x : 2 2 u↔ { x,v }: 2 x↔ { u,w }: 2 2 2 v→w : 2 v→w : 2 2 v→w : 2 v w v w v↔ { u,w }: 2 w↔ { v,x }: 2 Setting in our example: Abstraction intuition • • Every node supports 2 wavelengths Every node sends 2 unit flows to neighbors • • Every edge supports 2 wavelengths Every edge has a capacity of 2 • • u wants to send traffic to x u still wants to send traffic to x • • v wants to send traffic to w v still wants to send traffic to w OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 8

Intuition for the Abstraction 2 u x u x u→x : 2 u x u→x : 2 u→x : 2 2 2 u↔ { x,v }: 2 x↔ { u,w }: 2 2 2 2 2 2 2 v→w : 2 v→w : 2 2 2 v→w : 2 v w v w v w v↔ { u,w }: 2 w↔ { v,x }: 2 2 Setting in our example: Abstraction intuition Result: • • • Every node supports 2 wavelengths Every node sends 2 unit flows to neighbors 2 fake flows between u,v & x,w • • • Every edge supports 2 wavelengths Every edge has a capacity of 2 No capacity left between u,v & x,w • • • u wants to send traffic to x u still wants to send traffic to x 2 units of capacity for u,v & x,w • • • v wants to send traffic to w v still wants to send traffic to w Real throughput of: 2 + 2 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 8

2 u x 0 0 2 Intuition for the Abstraction v w 2 u x u x u→x : 2 u x u→x : 2 u→x : 2 2 2 u↔ { x,v }: 2 x↔ { u,w }: 2 2 2 2 2 2 2 v→w : 2 v→w : 2 2 2 v→w : 2 v w v w v w v↔ { u,w }: 2 w↔ { v,x }: 2 2 Setting in our example: Abstraction intuition Result: • • • Every node supports 2 wavelengths Every node sends 2 unit flows to neighbors 2 fake flows between u,v & x,w • • • Every edge supports 2 wavelengths Every edge has a capacity of 2 No capacity left between u,v & x,w • • • u wants to send traffic to x u still wants to send traffic to x 2 units of capacity for u,v & x,w • • • v wants to send traffic to w v still wants to send traffic to w Real throughput of: 2 + 2 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 8

2 u x TE performs TE+TP on abstraction details in the paper 0 0 2 Intuition for the Abstraction v w 2 u x u x u→x : 2 u x u→x : 2 u→x : 2 2 2 u↔ { x,v }: 2 x↔ { u,w }: 2 2 2 2 2 2 2 v→w : 2 v→w : 2 2 2 v→w : 2 v w v w v w v↔ { u,w }: 2 w↔ { v,x }: 2 2 Setting in our example: Abstraction intuition Result: • • • Every node supports 2 wavelengths Every node sends 2 unit flows to neighbors 2 fake flows between u,v & x,w • • • Every edge supports 2 wavelengths Every edge has a capacity of 2 No capacity left between u,v & x,w • • • u wants to send traffic to x u still wants to send traffic to x 2 units of capacity for u,v & x,w • • • v wants to send traffic to w v still wants to send traffic to w Real throughput of: 2 + 2 OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 8

Takeaway • TE performs TE and TP due to the abstraction ◦ Details in the paper • Consistent update methods for flows carry over ◦ Abstraction enables cross-layer updates for free • Support for major TE types (max. throughput, k-shortest path routing etc.) OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 9

Testbed: Demonstration of TP in Practice Physical setup of our testbed Logical setup OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 10

Experiment: Demonstration of TP in Practice ◦ Traffic from A to C (via A-C and A-B-D-C) Experiment setup Logical setup OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 11

Experiment: Demonstration of TP in Practice ◦ Traffic from A to C (via A-C and A-B-D-C) ◦ Fail A-B link Experiment setup Logical setup OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 12

Experiment: Demonstration of TP in Practice ◦ Traffic from A to C (via A-C and A-B-D-C) ◦ Fail A-B link ◦ Controller notices cut & shifts wavelength (s) Experiment setup Logical setup OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 13

Simulations: k-Shortest Path Routing OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 14

Simulations: k-Shortest Path Routing • Comparison: Standard ILP (JointOpt) vs. our approach (OptFlow) OptFlow: A Flow-based Abstraction for Programmable Topologies (SOSR'20) 14