Transcript new-bergamasco-bcn-july-plenary-0705

Updates on Backward
Congestion Notification
Davide Bergamasco ([email protected])
Cisco Systems, Inc.
IEEE 802 Plenary Meeting
San Francisco, USA
July 20, 2005
1
Agenda
• Previous presentation
• May 2005 IEEE 802.1 Interim Meeting in Berlin, Germany
• http://www.ieee802.org/1/files/public/docs2005/newbergamasco-backward-congestion-notification-0505.pdf
• Updates
• Algorithm
• Derivative to improve stability
• Solicit Bit to accelerate recovery
• AQM in rate limiter queues to reduce blocking
• Simulations
2
Queue Stability
Q
• ISSUE: Overshoots
Stop Generation of BCN Messages
+
-
-
+
+
-
-
and undershoots
accumulate over time
• SOLUTION: Signal
+
only when
• Q > Qeq && dQ/dt > 0
• Q < Qeq && dQ/dt < 0
• Easy to implement in
Qeq
hardware: just an
Up/Down counter
• Increment @ every
enqueue
• Decrement @ every
dequeue
t
• Reduces signaling
rate by 50%!!
3
Solicit Bit
• ISSUE: When the rate
is very low, recovery
may take too long
because of sampling.
R
• SOLUTION: Solicit
Bit in RL tag
• if R < Rsolicit Solicit
Rsolicit
bit is set
• if R >= Rsolicit Solicit
bit is cleared
Rmin
BCN0
t
Random
Time
BCN+2 BCN+2 BCN+4
Force Bit
On
Force Bit
Off
BCN+1
• If possible, CP will
generate a BCN+ for
every frame with
Solicit bit on,
regardless of
sampling
4
Changes to Detection & Signaling
T-4
T-3
T-2
FULL QUEUE
T-1
T+0
T+1
T+2
T+3
T+4
EMPTY QUEUE
EQUILIBRIUM
IN
OUT
Sample
Frame with
Probability P
MESSAGE TO GENERATE
Sampled
Frame?
Yes
No
RL
Tagged
Frame?
No
BCN 0
BCN-4
BCN-3
BCN-2
BCN-1
NOP
No Message
No
Yes
MESSAGE TO GENERATE
BCN 0
BCN-4
BCN-3
BCN-2
BCN-1
No
Message
BCN+1
BCN+2
BCN+3
BCN+4
dQ/dt < 0?
0
BCN
type
-
MESSAGE TO GENERATE
RL Tag
&& Solicit
Bit Set?
Yes
+
Send
BCN
dQ/dt > 0
Yes
No
Yes
No Message
BCN+1
BCN+2
BCN+3
BCN+4
NOP
No
NOP
5
Rate Limiter Queue AQM
F1
R1
F2
R2
Fn
No
Match
• ISSUE: Blocking @
Rn
RL queues due to
buffer exhaustion
• SOLUTION: add an
Flow Control
Data IN
Tail Drop
Packets Marked with
RATE_LIMITED_TAG
Data OUT
Control IN
EDGE
NODE
AQM mechanism to
control buffer usage
BCN Messages
from congested
point
NETWORK
CORE
6
Rate Limiter Queue AQM
• Traditional AQM such as RED (mark/drop) don’t work well for RL queues:
• Buffer too small
• Very few flows
• Traffic statistics very different from Internet traffic
• A novel and very simple solution based on:
• Threshold on the RL queue QAQM (e.g., 10 pkts)
• Fixed drop or mark probability P (e.g., 1%)
• Two counters:
•
CTCP: Number of TCP packets in the RL queue
• CUDP: Number of UDP packets in the RL queue
Drop or mark TCP packets with probability P when CTCP > QAQM
•
• Drop UDP packets when CUDP > QAQM
QAQM
7
Simulation Environment (1)
ES6
Core Switch
TCP Bulk
SJ
TCP On/Off
Congestion
DR2
TCP Ref1
TCP Ref2
ES1
ES2
ES3
STb2 STo2
STb3 STo3
ES4
ES5
SR2
SR1
STb1 STo1
STb4 STo4
DTb
DTo
DR1
8
Simulation Environment (2)
• Short Range, High Speed DC Network
•
•
•
•
Link Capacity = 10 Gbps
Switch latency = 1 s
Link Length = 100 m (.5  s propagation delay)
Control loop delay ~ 3 s
• Workload
1) TCP only
•
STb1-STb4: 3 parallel connections transferring 1 MB each continuosly
•
STi1-STi4: 3 parallel connections transferring 1 MB then waiting 10 ms
•
SR1: 1 connection transferring 10 KB (avg 16 s wait)
•
SR2: 1 connection transferring 10 KB (1s wait)
2) 80% TCP + 20% UDP
•
STb1-STb4: same as above
•
STi1-STi4: same as above
•
SR1-SR2: same as above
•
SU1-SU4: variable length bursts with average offered load of 2 Gbps
9
Simulation Goals
• Study the performance of BCN with various
congestion management techniques at the RL
• No Link-level Flow Control
• Link-level Flow Control
• Link-level Flow Control + RL simple AQM (drop/mark)
• Metrics:
• Throughput and Latency of TCP bulk and on/off
connections
• Throughput and Latency of Reference Flows
• Bottleneck Link Utilization
• Buffer Utilization
10
Bulk & On/Off Application Throughput &
Latency (Workload 1: TCP Only)
RL Congestion
Management
Mechanism
Bulk TCP
Throughput
(Tps)
Bulk TCP
Latency
(s)
No Flow
Control
67.17
15,220
25.92
27,880
9.85
Flow Control
63.00
15,970
32.92
20,337
10.00
Flow Control +
RL AQM (drop)
61.83
16,249
33.42
19,570
10.00
Flow Control +
RL AQM (mark)
59.17
17,043
36.67
16,873
10.00
Best
On/Off TCP
Throughput
(Tps)
On/Off
TCP
Latency
(s)
Throughput
on
Bottleneck
link (Gbps)
Worst
11
Reference Applications Throughput &
Latency (Workload 1: TCP Only)
RL Congestion
Management
Mechanism
Ref1 TCP
Throughput
(Tps)
Ref1 TCP
Latency
(s)
Ref2 TCP
Throughput
(Tps)
Ref2 TCP
Latency
(s)
No Flow
Control
6702
132.8
30334
31.97
Flow Control
7108
124.30
31038
31.22
Flow Control +
AQM (drop)
7210
122.33
31307
30.94
Flow Control +
AQM (mark)
7419
119.21
31362
30.89
Best
Worst
12
Buffer Utilization: No FC
13
Buffer Utilization: FC
14
Buffer Utilization: FC + RL AQM (drop)
15
Buffer Utilization: FC + RL AQM (mark)
16
Summary & Next Steps
• A number of improvements have been
made to BCN
• Derivative to improve stability
• Solicit Bit to speed up recovery
• AQM in RL queues to reduce blocking
• Future Steps
• Build a Prototype???
•…
17
18