«Insomnia in the Access» Or How to Curb Access Network Related Energy Consumption Marco Canini EPFL Joint work with Eduard Goma, Alberto Lopez Toledo, Nikolaos Laoutaris (Telefonica Research), Dejan Kostić (EPFL), Pablo Rodriguez (Telefonica Research) , Rade Stanojević (IMDEA Networks), Pablo Yagüe Valentín (Telefonica Research)
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
«Insomnia in the Access»Or How to Curb Access Network Related Energy Consumption
Marco CaniniEPFL
Joint work with
Eduard Goma, Alberto Lopez Toledo, Nikolaos Laoutaris (Telefonica Research), Dejan Kostić (EPFL), Pablo Rodriguez (Telefonica Research) ,
Rade Stanojević (IMDEA Networks), Pablo Yagüe Valentín (Telefonica Research)
2
Greening the ICT
Datacenters Access network
?
BackbonePCs &
peripherals
3
Energy consumption of telcos
2.1 TWh
3.7 TWh
4.5 TWh
9.9 TWh
4
Access dominates energy consumption
20-30% 70-80% ACCESSBackbone/Metro/Transport
5
A typical DSL access network
CORE METRO
Gateway
USER PARTISP PARTACCESS
6
A typical DSL access network
CORE METRO
USER PARTISP PARTACCESS
DSL Access Multiplexers(DSLAMs)
Cable bundle
Central Office
Gateway
Distribution frames of a
Central Office7
Can we save energy?How?
1. Challenges in greening the access2. Two practical techniques can save of
the access energy …3. With a performance bonus
8
9
WHY DOES THE ACCESS CONSUME SO MUCH?
10
Individually, they do not consume a lot#1: Huge number of devices
PhotoBlackburn
But collectively …
2 orders of magnitude more gateways than DSLAMs
1 order of magnitude more DSLAMs than metro devices
2 orders of magnitude more DSLAMs than backbone devices
11
2#: High per bit energy consumption
At full load, access devices 2-3 orders of magnitude higher than metro/backbone
ACCESSBackbone/Metro
12
Energy proportionality?
0 10 20 30 40 50 60 70 80 90 1000
10
20
30
40
50
60
70
80
90
100
Power Energy efficiency
Utilization [%]
Pow
er u
sage
[% o
f pea
k]
100%70%
13
3#: Utilization < 10%
0 10 20 30 40 50 60 70 80 90 1000
10
20
30
40
50
60
70
80
90
100
Power Energy efficiency
Utilization [%]
Pow
er u
sage
[% o
f pea
k]
0 2 4 6 8 10 12 14 16 18 20 220%
2%
4%
6%
8%
10%
Daily utilization of 10K access links in a commercial ADSL provider
uplinkdownlink
Time [h]
Aver
age
utiliz
ation
[%]
… but most of the time hereAlready bad here
14
Sleeping saves energy
Sleep-on-Idle (SoI)Devices enter sleep mode upon periods of inactivity
15
METRO
SoI fails in access networks
USER PARTISP PARTACCESS
An ADSL line needs 1 minute to wake up… but cannot enjoy a minute’s sleep
16
What if we can put 80% of gateways to sleep?
15 W
100 W
1 W per modem
Save big fraction at the user side ISPs … not so much
17
Line cards very unlikely to sleep by SoI
Line cardsModem off
DSLA
M
Modem on
Static assignment of lines to DSLAM ports is a problem
⟹Greening the user part: aggregation
⟹Greening the ISP part: line switching
OUR APPROACH
18
19
Greening the user part – Aggregation
On average 5-6 WiFi networks overlap in typical urban settings
?
???
Broadband Hitch-Hiking (BH2)Threshold-based heuristic algorithm: direct traffic to
neighbor gateways during light traffic conditions
20
Broadband Hitch-Hiking(BH2)
5% 30%
20%
25%
0% 35%5%
30%50%45%
Load on home gateway is low direct light traffic to a neighbor gateway and let home gateway sleepLoad on neighbor gateway is low look for another neighbor or go back to home gatewayLoad on neighbor gateway is high go back to home gateway
15 W
21
BH2 under the hoodGateways assumed to have:• SoI• Ability to wake on traffic
BH2 terminals use WiFi card virtualization:[Giustiniano et al., MobiCom ’10; Kandula et al., NSDI ’08]
• Estimate load on all gateways in range• Maintain connectivity with 1 or few
backup gateways
Only modify the terminal wireless card driver
22
Greening the ISP part – Line switchingDS
LAM
Line cards
AT&T operators connecting transatlantic calls at the international switchboard in New York, circa 1930. (www.corp.att.com)
40-way switch
Full switching maximizes savings …but cost quickly grows with the number of ways
Switching is for lines, not for packets
23
Small 4-way switches are enough
Each k-switch packs active lines to the top
DSLA
M
Line cards
Put line cards to sleepSimple micro-electro-mechanical switches with near-zero power consumption
4-way switches
Evaluation
• Traces: CRAWDAD UCSD – 272 clients, 40 gateways• Scenario: 4 x 12-port line cards; 5.6 avg. gateways• Results for 4 schemes• Baseline: no sleep
– Gateways, modems, line cards don’t sleep– Each terminal only connects to its home gateway
24
25
How many gateways can sleep?
0 5 10 15 20 240
10
20
30
40
Time [h]Num
ber o
f sle
epin
g ga
tew
ays
SoI
Scheme 1: SoI for gateways, modems & line cards
Peak hours
26
How many gateways can sleep?
0 5 10 15 20 240
10
20
30
40
Time [h]Num
ber o
f sle
epin
g ga
tew
ays
SoI
BH2
BH2 puts to sleep 70-90% of gateways
Scheme 2: BH2 + k-switchBH2 every 150 s + 12 x 4-switches
27
How many gateways can sleep?
0 5 10 15 20 240
10
20
30
40
Time [h]Num
ber o
f sle
epin
g ga
tew
ays
SoI
BH2
Optimal
Scheme 3: OptimalBest clients to gateways assignment + full switch
28
What is the impact of gateway density?
1 2 3 4 5 6 7 8 9 100
5
10
15
20
25
30
35
40
Mean number of available gateways per clientNum
ber o
f sle
epin
g ga
tew
ays
Just home + 2 neighbors
50%+ of gateways sleep
29
How much energy can we save?
0 5 10 15 20 240
20
40
60
80
100
Time [h]Ene
rgy
savi
ngs
vs n
o-sl
eep
[%]
SoI
Optimal
BH2 + k-switch
BH2 + k-switch saves 66%Optimal savings are 80%
30
What are the savings for the ISP?
0 5 10 15 20 240
10
20
30
40
50
60
Time [h]
% o
f tot
al e
nerg
y sa
ving
son
the
ISP
sid
e
SoI + k-switch
SoI
Scheme 4: SoI + k-switchCombine switching with SoI
31
What are the savings for the ISP?
0 5 10 15 20 240
10
20
30
40
50
60
Time [h]
% o
f tot
al e
nerg
y sa
ving
son
the
ISP
sid
e
SoI + k-switch
SoI
BH2 + k-switch
Savings come from both aggregation and switching
32
What are the savings for the ISP?
0 5 10 15 20 240
10
20
30
40
50
60
Time [h]
% o
f tot
al e
nerg
y sa
ving
son
the
ISP
sid
e
Optimal SoI + k-switch
SoI
BH2 + k-switch
Savings come from both aggregation and switching
33
Prototypedeployment
2 4 6 8 10 12 14 16 18 20 22 24 26 28 300
2
4
6
8
Time [mins]Num
ber o
f sle
epin
g A
Ps
SoIBH2
1st floor 2nd floor 3rd floor
+
Limitation: home gateway + 2 neighbor gateways per client
34
Incentives
SecurityPrivacy
Considerations for deployment
35
A performance bonus
0 2 4 6 8 10 12 16 20
0
10
20
30
40
50
Number of inactive lines
Avg
. spe
edup
[%]
62 Mbps; loop lengths 50-600 m
Powering off lines makes the remaining … go faster due to reduced crosstalk!
Bonus: reduced crosstalk
Conclusions
• 80% energy can be saved at the access• Aggregation + switching save 66% of energy• Surprising result: turning DSL modems off increases
the performance of remaining modems
• Applying our solution to all DSL users worldwide, yields savings of 33 TWh per year– ½ of energy going into US datacenters Thank you!