Title: Reducing Network Energy Consumption via Sleeping and Rate Adaptation
1Reducing Network Energy Consumption via Sleeping
and Rate Adaptation
2Reducing Network Energy Consumption via Sleeping
and Rate Adaptation
- Authors
- Sergiu Nedevschi
- UC Berkeley Intel Research
- Lucian Popa (UC Berkeley)
- Sylvia Ratnasamy (Intel Research)
- Gianluca Iannaccone (Intel Research)
- David Wetherall (U Washington Intel Research)
- My Name Anand Seetharam
-
3Motivation
- Network energy consumption a growing issue
- Equipment increasingly power-hungry (power
density) - Rising energy costs (significant fraction of TCO)
- Environmental concerns
- Energy Efficient Ethernet Taskforce (IEEE 802.3
az) - Focuses on saving network energy for Ethernet
-
-
4Opportunity
- Networks are provisioned for peak-load
- phone network needs to work on 1st JAN, at 12AM
- Average utilization is low
Network Utilization
ATT switched voice 33
Internet Links 15
Private line networks 3-5
LANs 1
Data networks are lightly utilized, and will
stay that way A. M. Odlyzko, Review of Network
Economics, 2003
5Opportunity
- Energy consumption proportional to capacity, not
actual utilization!! - Idle energy consumption is high
- For example, a Cisco GSR linecard draws
- Chabarek etal, INFOCOM08
- 80W idle
- 90W fully loaded
Most energy consumed by networks is wasted!
Goal Make network energy consumption reflect
utilization levels, not peak provisioning
6Idea
- Key Idea Let network equipment sleep for brief
periods or slow down when lightly loaded to save
energy - Inspiration Use of sleep and performance
states in PCs, processors - Rationale E pidle Tidle pactive
Tactive - Assumptions We assume support for
sleep/performance states in NICs, linecards,
switches, and routers and consider how to best
use them - Depend on
- Type/extent of hardware support for sleep and
performance states - Careful use of these states to protect
performance and maximize savings
7Outline
- Key questions and method
- Sleeping
- Rate adaptation (slowing down)
- Sleep vs. Rate adaptation
81. Key questions and method
- How much energy can we save without compromising
performance? - Can we realize these savings with practical
schemes? - Methodology
- Model hardware support for sleep and rate
adaptation - Evaluate savings/performance with simulations
(ns) - Abilene and Intel topologies and their traffic
workloads - Look for (unrealistic) bounds as well as
practical schemes
92. Sleeping states
- Model
- Single sleep state with power psleepltlt pidle
- d transition period (ms)
- Timer or activity-driven wakeup
- Interfaces sleep independently
- Metrics
- Energy savings in time asleep
- Performance in loss and max delay
power
pidle
(idle)
(sleep)
psleep
time
d
10When can a link sleep?
- Packets over a link
- sleep time depends on
- Buffer and burst
time
2
3
4
6
1
5
7
Transition time
d
11Making sleep gaps on links with buffer burst
(BB)
- Basic idea use limited buffering at ingress to
create predictable and useful sleep gaps (gt2d)
do once, adds bounded delay
5ms
20ms
2ms
R1
R2
R3
_at_ t8 tB8 t2B8
_at_ t28 tB28 t2B28
tx _at_ t1 tB1 t2B1
wake _at_ t3 tB3
t2B3
12Coordination among ingresses
- Basic idea align bursts/gaps on links in
networks by adjusting relative timing phase of
different ingresses
t, tB,
I1
8ms
coordinate burst times to align in the network
R
3ms
t5, t5B,
I2
13Potential for savings with sleep (optBB)
- perfect coordination not generally possible
t1
1ms
t1 1ms t2 20ms
I1
R1
15ms
20ms
t2
t1 15ms t2 2ms
I1
R2
2ms
- Upper bound (optBB) Global search to find
ingress transmission times that maximize
network-wide sleep
14Potential benefits of sleeping
Abilene, transition time1ms, B10ms
idle (bound) WoA (pareto) WoA (CBR) optBB(CBR)
Upper bound for any scheme
Upper bound without buffering/shaping
Upper bound with buffering/shaping
A little shaping can get most of the utilization
gain
15Practical sleeping algorithm (practBB)
- Ingress buffers and transmits packets in a bunch
every Bms - Within bunch, packets are organized by egress
- Router interfaces wake to process bursts
- Router interfaces sleep if start of next burst is
gt2d ms away
16No coordination (practBB)
Abilene, transition time1ms, B10ms
Practical algorithm realizes most of the benefit
17Impact of sleeping on delay
Abilene, transition time1ms
98th percentile delay (ms)
No added loss added delay bounded by Bms
18Impact of sleep Any Losses?
- No additional losses are incurred until
utilizations come close to saturating some links. - Losses greater than 0.1 occur at
Scheme Utilization
Default 41
B 10ms 38
B 25ms 36
Abilene, network utilization5
19Impact of sleep transition time
Quick transitions (preferably lt 1ms) needed
Abilene, network utilization5
20Outline
- Key questions and method
- Sleeping
- Rate adaptation (slowing down)
- Sleep vs. Rate adaptation
213. Rate adaptation states
- Model
- N performance states
- Rates r1, , rn and pi lt pi1
- d transition period (ms)
- Interfaces can rate-adapt independently
- Metrics
- Energy savings in average rate reduction
- Performance in loss and max delay
power
(1G)
pi1
(100M)
pi
time
d
22Using performance states
- Basic idea decrease rate as much as possible
without introducing more than than d ms per hop
- Optimal algorithm ideal service curve
follows shortest Euclidean distance.
23Practical rate adaptation (practRA)
- Idea lower rate if doing so will maintain
minimal queuing delay (of at most d ms)
aggressively increase rate to clear buildup
- Algorithm
- rf estimated arrival rate as average (EWMA) of
past arrivals - q current queue size
- d target maximum queuing delay
- ri current link operating rate
- Rules
- increase to ri1 iff (q/ri gt d) OR (drf q)/ri1gt
(d- d) - decrease to ri-1 iff (q 0) AND (rf lt ri-1 )
- duration since last rate change gt k d (k4)
Leave headroom for transition time
Avoid frequent changes
24Benefits of rate adaptation
Abilene, transition time d 1ms, d3ms
- Added delay lt d (hops)
- No observed packet loss
25Outline
- Key questions and method
- Sleeping
- Rate adaptation (slowing down)
- Sleep vs. Rate adaptation
26Models of future power profiles
Fraction of power that doesnt scale with rate
- pactive C fn(rate)
-
- pidle C ß fn(rate)
- psleep µ pidle(rmax)
Rate scaling function
fn(rate) rate frequency scaling
fn(rate) rate3 dynamic voltage scaling
Idle/Active Workload Ratio
Power reduction using sleep
27Sleeping and rate adaptation (DVS-r3)
28Sleeping and rate adaptation (Frequency Scaling
-r)
29Observations
- The authors say
- Hence to avoid complex interactions, we consider
that the - whole network , or at least well-defined
components of it, run - either rate adaption or sleep
- But both schemes can be combined to give better
results. - For eg In rate adaptation one can try to put the
links to sleep - instead of keeping them in the idle
state.
30Observations
- When rate adaptation is done using frequency
scaling the authors themselves - say that for values (C0.3 and ß 0.1) and (C0.3
and ß 0.8) the savings - obtained are poor and add little additional
information. -
- My observation is that rate adaptation (frequency
scaling) gives - no gain in terms of energy.
31Thank you. Questions?