Slide 1

Slide 1 text

Greening Backbone Networks Emaad Ahmed Manzoor CSE691: Green IT & Smart Grid March 2016 Will Fisher, Martin Suchara, Jennifer Rexford First Workshop on Green Networking, SIGCOMM ’10

Slide 2

Slide 2 text

Backbone Networks

Slide 3

Slide 3 text

https://en.wikipedia.org/wiki/File:Internet_map_1024.jpg

Slide 4

Slide 4 text

https://commons.wikimedia.org/wiki/File:NSFNET-traffic-visualization-1991.jpg

Slide 5

Slide 5 text

Aggregate Links (2-40 per-cable)

Slide 6

Slide 6 text

http://www.klalaboratories.com/service/asset/network_cabling_3_lg.jpg

Slide 7

Slide 7 text

https://www.reddit.com/r/mildlyinteresting/comments/3w14gh/this_giant_cable_stuffed_with_hundreds_of_wires/

Slide 8

Slide 8 text

Router Line Cards

Slide 9

Slide 9 text

https://farm4.staticflickr.com/3572/3567497578_aa942f35dd_b.jpg

Slide 10

Slide 10 text

Not power-proportional https://farm4.staticflickr.com/3572/3567497578_aa942f35dd_b.jpg

Slide 11

Slide 11 text

Power-proportional Router: Architectural Design and Experimental Evaluation, Bin Liu et. al. IWQoS 2014. Not power-proportional

Slide 12

Slide 12 text

Cannot be put to “sleep”* Not power-proportional https://farm4.staticflickr.com/3572/3567497578_aa942f35dd_b.jpg

Slide 13

Slide 13 text

Cannot be put to “sleep”* Not power-proportional *(2010) https://farm4.staticflickr.com/3572/3567497578_aa942f35dd_b.jpg

Slide 14

Slide 14 text

Average link-utilisation in backbone networks is 30-40%.

Slide 15

Slide 15 text

Can we power-down links (line cards) and adjust data flow based on traffic demand?

Slide 16

Slide 16 text

Problem formalisation Naive solution Integer linear program Efficient heuristics Evaluation Future

Slide 17

Slide 17 text

Problem formalisation Naive solution Integer linear program Efficient heuristics Evaluation Future

Slide 18

Slide 18 text

Input

Slide 19

Slide 19 text

G(V, E) s t w u v c(u,t) c(w,v) c(v,t) c(s,w) c(s,u) Topology Input

Slide 20

Slide 20 text

G(V, E) s t w u v c(u,t) c(w,v) c(v,t) c(s,w) c(s,u) Topology D (s1, t1, h2) (s2, t2, h2) . . . Demands Input

Slide 21

Slide 21 text

G(V, E) s t w u v c(u,t) c(w,v) c(v,t) c(s,w) c(s,u) Topology D (s1, t1, h2) (s2, t2, h2) . . . Demands Cables per link B Input

Slide 22

Slide 22 text

Output

Slide 23

Slide 23 text

s t w u v c(u,t) c(w,v) c(v,t) c(s,w) c(s,u) Flow of demand d on each link fd(s,u) fd(s,w) fd(w,v) fd(v,t) fd(u,t) fd(u, v) Output

Slide 24

Slide 24 text

s t w u v c(u,t) c(w,v) c(v,t) c(s,w) c(s,u) Flow of demand d on each link fd(s,u) fd(s,w) fd(w,v) fd(v,t) fd(u,t) Total flow on link (u,v) f(u, v) = X d fd(u, v) fd(u, v) Output

Slide 25

Slide 25 text

Problem formalisation Naive solution Integer linear program Efficient heuristics Evaluation Future

Slide 26

Slide 26 text

min number of powered cables s.t. link loads ≤ capacities flow conservation carries all traffic demands

Slide 27

Slide 27 text

min s.t. link loads ≤ capacities flow conservation carries all traffic demands X (u,v)2E f(u, v)

Slide 28

Slide 28 text

min s.t. flow conservation carries all traffic demands X (u,v)2E f(u, v) X D fd(u, v)  c(u, v), 8(u, v) 2 E

Slide 29

Slide 29 text

min s.t. carries all traffic demands X (u,v)2E f(u, v) X D fd(u, v)  c(u, v), 8(u, v) 2 E X v2V fd(u, v) = X w2V fd(w, u), 8d, 8u 6= sd, td

Slide 30

Slide 30 text

min s.t. X (u,v)2E f(u, v) X D fd(u, v)  c(u, v), 8(u, v) 2 E X v2V fd(sd, v) = X w2V fd(w, td) = hd, 8d X v2V fd(u, v) = X w2V fd(w, u), 8d, 8u 6= sd, td

Slide 31

Slide 31 text

min number of powered cables s.t. link loads ≤ capacities flow conservation carries all traffic demands Models what if links are energy proportional? • Fractional linear program • Bounds on power savings

Slide 32

Slide 32 text

Upper bound on energy savings f(u,v) = 6.3 Gbps (solution of FLP) c(u,v) = 10 Gbps (given topology) B = 10 (given cables per link) — Round down: 6 Gbps — Turn off: 4 cables

Slide 33

Slide 33 text

Lower bound on energy savings f(u,v) = 6.3 Gbps (solution of FLP) c(u,v) = 10 Gbps (given topology) B = 10 (given cables per link) — Round up: 7 Gbps — Turn off: 3 cables

Slide 34

Slide 34 text

Unfortunately, the lower bound solution is feasible but can be extremely suboptimal.

Slide 35

Slide 35 text

(S, Ti, ✏) (Bi, Bi+1, ✏)

Slide 36

Slide 36 text

Problem formalisation Naive solution Integer linear program Efficient heuristics Evaluation Future

Slide 37

Slide 37 text

The bundle size B needs to be explicitly taken into account.

Slide 38

Slide 38 text

min X (u,v)2E nuv Minimise the sum of the number of active cables on each link f(u, v)  nuv B c(u, v), 8(u, v) 2 E Number of active cables need to be sufficient for the demand

Slide 39

Slide 39 text

min X (u,v)2E nuv Minimise the sum of the number of active cables on each link f(u, v)  nuv B c(u, v), 8(u, v) 2 E Number of active cables need to be sufficient for the demand NP-Complete — D2CIF

Slide 40

Slide 40 text

Problem formalisation Naive solution Integer linear program Efficient heuristics Evaluation Future

Slide 41

Slide 41 text

Fast Greedy Heuristic 1. Solve FLP - Obtain f(u,v), round up. 2. Find (u,v) with the most unused capacity - Remove one cable from the link. - Solve the FLP with the new capacities. Feasible? Remove a cable from the link (u,v).

Slide 42

Slide 42 text

Fast Greedy Heuristic 1. Solve FLP - Obtain f(u,v), round up. 2. Find (u,v) with the most unused capacity - Remove one cable from the link. - Solve the FLP with the new capacities. Feasible? Remove a cable from the link (u,v). Not feasible? Mark link (u,v) as final.

Slide 43

Slide 43 text

Fast Greedy Heuristic 1. Solve FLP - Obtain f(u,v), round up. 2. Find (u,v) with the most unused capacity - Remove one cable from the link. - Solve the FLP with the new capacities. Feasible? Remove a cable from the link (u,v). Not feasible? Mark link (u,v) as final. Repeat from 2, ignoring all final links

Slide 44

Slide 44 text

Exhaustive Greedy Heuristic Remove the cable that - Retains solution feasibility. - Minimises the marginal penalty of each removal. Marginal penalty - Increases as excess traffic is rerouted along longer paths. - Measures as the FLP objective difference after cable removal.

Slide 45

Slide 45 text

Bi-level Greedy Heuristic Exhaustive greedy heuristic with pairs of cables.

Slide 46

Slide 46 text

Problem formalisation Naive solution Integer linear program Efficient heuristics Evaluation Future

Slide 47

Slide 47 text

FGH EGH BGH

Slide 48

Slide 48 text

Abilene Topology

Slide 49

Slide 49 text

No content

Slide 50

Slide 50 text

No content

Slide 51

Slide 51 text

No content

Slide 52

Slide 52 text

54% savings with B = 1 73% savings with B = 3

Slide 53

Slide 53 text

No content

Slide 54

Slide 54 text

Problem formalisation Naive solution Integer linear program Efficient heuristics Evaluation Future

Slide 55

Slide 55 text

https://tools.ietf.org/html/draft-zhang-greennet-01 Impact

Slide 56

Slide 56 text

Dynamic demands Real-world traffic traces Approximation algorithms (?) Routing constraints

Slide 57

Slide 57 text

.