Circuit switching in the internet l.jpg
This presentation is the property of its rightful owner.
Sponsored Links
1 / 50

Circuit switching in the Internet PowerPoint PPT Presentation


  • 64 Views
  • Uploaded on
  • Presentation posted in: General

Circuit switching in the Internet. Ph.D. oral examination Pablo Molinero-Fern á ndez 21 st May 2002. GAP OF 5x!!. Traffic doubles every year. Cost and complexity of five times as many central offices is prohibitive. Optics removes bandwidth constraints. But we cannot buffer light!.

Download Presentation

Circuit switching in the Internet

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript


Circuit switching in the internet l.jpg

Circuit switching in the Internet

Ph.D. oral examination

Pablo Molinero-Fernández

21st May 2002


Traffic doubles every year l.jpg

GAP OF 5x!!

Traffic doubles every year

Cost and complexity of five times as many

central offices is prohibitive


Optics removes bandwidth constraints l.jpg

Optics removes bandwidth constraints

But we cannot buffer light!


Circuit switching l.jpg

Circuit switching

  • Link bandwidth is reserved

  • Need signaling

  • No buffering

  • No processing in data path


Packet switching l.jpg

Packet switching

  • Link bandwidth is shared

  • Buffering

  • Per-packet processing


Contributions l.jpg

Contributions

The Internet needs circuit switching

in the core

TCP Switching: how to integrate circuit switching in the core

Provisioning of fat pipes in an all-optical backbone


How we think the internet is l.jpg

How we think the Internet is


Why the internet uses packet switching l.jpg

Why the Internet usespacket switching

  • Efficient use of expensive links:

    • “Circuit switching is rarely used for data networks, ... because of very inefficient use of the links” – Bertsekas & Gallager ‘92

  • Resilience to failure of links & routers:

    • ”For high reliability, ... [the Internet] was to be a datagram subnet, so if some lines and [routers] were destroyed, messages could be ... rerouted” – Tanenbaum ‘96


What the internet is really like today l.jpg

What the Internet is really like today

SONET/SDH

DWDM


What the internet is really like today10 l.jpg

What the Internet is really like today

Modems, DSL


Performance criteria l.jpg

Performance criteria

Users

  • Response time

  • Service Level Agreement (QoS)

Carriers

  • Cost: Bandwidth efficiency

  • Reliability and stability

  • Low complexity


What users want l.jpg

What users want

Users

  • Response time

  • Service Level Agreement (QoS)

Carriers

  • Cost: Bandwidth efficiency

  • Reliability and stability

  • Low complexity


Response time of packets and circuits l.jpg

100 clients

1 server

1 Gb/s

Circuit sw

Packet sw

All but one of circuits finish earlier

Flow bandwidth

1 Gb/s

10 Mb/s

Avg response time

0.51 s

1 s

Max response time

1 s

1 s

Response time of packets and circuits

File = 10Mbit

x 100


Response time when blocking occurs l.jpg

100 clients

1 server

1 Gb/s

Circuit sw

Packet sw

A big flow can kill CS if it blocks the link

Flow bandwidth

1 Gb/s

10Mb/s;1Gb/s

Avg response time

10.50 s

1.10 sec

Max response time

10.99 s

10.99 sec

Response time when blocking occurs

File = 10Gbit/

10Mbit

x 99


Response time with flow rate limits l.jpg

100 clients

1 server

1 Gb/s

1 Mb/s

No difference between CS and PS in core

Circuit sw

Packet sw

Flow bandwidth

1 Mb/s

1 Mb/s

Avg response time

109.9s

109.9sec

Max response time

10,000 s

10,000 sec

Response time with flow rate limits

File = 10Gbit/

10Mbit

x 99


Analytical modeling l.jpg

Analytical modeling

  • Fluid model:

  • Many independent arrivals => Poisson

  • Service policies:

    • Packets: Processor Sharing

    • Circuits: FCFS

  • Service time distribution:

    • Flow size variance: Bimodal

    • Realistic flow size distrib: Pareto


Fluid model m bimodal n l.jpg

+

Flow size variance

Fluid model: M/BiModal/N

  • Access link is as fast as the core link

  • Hogging by long flows


Fluid model m bimodal n18 l.jpg

+

Flow size variance

Fluid model: M/BiModal/N

  • Flow rate limited by access link (1/N)

  • Same response time regardless of flow size variance, if N is large

N=2k


Fluid model m pareto n l.jpg

Fluid model: M/Pareto/N

Link hogging becomes very bad with heavy-tailed traffic, if ratio N=1


Fluid model m pareto n20 l.jpg

512

Fluid model: M/Pareto/N

  • Response time similar to IP if core/access ratio N is large

  • Typically N >> 1,000

N =2k


Users see little difference in response time l.jpg

Users see little difference in response time

  • Simulation of full networks

  • N is large => same response time

Circuit

switching

Packet

switching


Service level agreements l.jpg

Service Level Agreements

  • Packet switching:

    • Algorithms (WFQ, DRR, …) => not used

    • Thus we must overprovisioning => used and it works

  • Circuit switching:

    • Simple QoS: guaranteed BW => no jitter


What carriers want l.jpg

What carriers want

Users

  • Response time

  • Service Level Agreement (QoS)

Carriers

  • Cost: Bandwidth efficiency

  • Reliability and stability

  • Low complexity


Cost bandwidth efficiency l.jpg

Cost: Bandwidth efficiency

  • Argument: packets share all link BW => statistical multiplexing gain => more throughput with bursty traffic

  • Reality:

    • Internet avg. link utilization: 5-20% [Coffman&Odlyzko’02]

    • Phone avg. link utilization: ~33% [Odlyzko’99]

    • There is a glut of BW in the core [WSJ’00]

  • Result:

    • Packets more efficient, but BW is no longer a scarce resource


Carriers peak allocate their network l.jpg

OC-48

OC-3

OC-12

Carriers peak allocate their network

  • When over-provisioning, link BW is virtually peak allocated

  • That is exactly what circuit switching does

Source: Chuck Fraleigh ‘02


Reliability and stability i l.jpg

Reliability and stability (I)

  • Argument: because of the state, rerouting a circuit is more costly than with packets

  • Reality:

    • Internet average availability: 1220 min/year down time [Labovitz’99]

    • Phone average availability: 5 min/year down time [Kuhn’97]


Reliability and stability ii l.jpg

Reliability and stability (II)

  • Reality (cont.):

    • IP recovers in about 3 min (median), sometimes it takes over 15 min [Labovitz’01]

    • SONET required to recover in less than 50 ms

  • Result:

    • No evidence packet switching is more robust


Low complexity i l.jpg

Low complexity (I)

  • Argument: No per-flow state => packet switching is simpler

  • Reality:

    • PS: 8M lines of code in core router [Cisco’s IOS ‘00]

    • CS: 18M lines of code in telephone switch [AT&T/Lucent 5ESS ‘96]

    • CS: 3M lines of code in transport switch [’01]

  • Result:

    • Packet switching does not seem inherently less complex than circuit switching


Functions in a packet switch l.jpg

Interconnect

Egress linecard

Ingress linecard

Route lookup

Buffer

ing

Framing

TTL process

ing

Framing

Buffer

ing

QoS schedul

ing

Interconnect scheduling

Control plane

Control path

Data path

Scheduling path

Functions in a packet switch


Functions in a circuit switch l.jpg

Ingress linecard

Egress linecard

Interconnect

Framing

Framing

Interconnect scheduling

Control plane

Control path

Data path

Functions in a circuit switch


Low complexity ii l.jpg

Low complexity (II)

  • Argument: IP does not have the signaling of circuits switches => Routers go faster

  • Reality:

    • IP does almost same operations on every packet as a circuit switch on the circuit establishment

    • CS has no work to do once circuit is established

  • Result:

    • The fastest commercially-available circuit switches [Ciena ’01, Lucent ‘01] have 5x the capacity of the fastest routers [Cisco ’01, Juniper ’02]


Network architecture l.jpg

Network architecture

  • Use packet switching

  • Better response time (ratio N small)

  • Efficient use of the spectrum

LANs &

wireless

  • If metro-to-access BW ratio (N) is small => use packets

  • Otherwise use what costs less

MANs

  • Use circuit switching

  • More capacity, reliability

  • Similar response time & QoS

WANs


Contributions33 l.jpg

Contributions

The Internet needs circuit switching

in the core

TCP Switching: how to integrate circuit switching in the core

Provisioning of fat pipes in an all-optical backbone


Integration of circuits and packets l.jpg

Integration of circuits and packets

  • Create a separate circuit for each user flow

  • IP controls circuits

  • Optimize for the most common case

    • TCP (90-95% of traffic)

    • Data (9 out of 10 pkts)

TCP

Switching


Tcp switching exposes circuits to ip l.jpg

TCP Switching exposes circuits to IP

IP routers

TCP Switches


Tcp creates a connection l.jpg

Destina-tion

Source

Router

Router

Router

SYN

SYN+ACK

DATA

Packets

Packets Packets

Packets

TCP “creates” a connection


Let tcp leave state behind l.jpg

Boundary TCP-SW

Core TCP-SW

Boundary TCP-SW

Destina-tion

Source

SYN

Create circuit

SYN+ACK

Create circuit

DATA

Packets

One Circuit

Packets

Let TCP leave state behind


What is a typical flow l.jpg

What is a typical flow?

  • Most traffic are TCP connections:

    • Taking less than 10 s, 12 packets and 4 KBytes

    • Obtaining less than 100 Kbps

    • ~40% of the flows continue transmitting ACKs after sending a FIN (asymmetrical closures)


State management feasibility l.jpg

State management feasibility

  • Amount of state

    • Minimum circuit = 56 Kb/s.

    • 178,000 circuits for OC-192.

  • Update rate

    • About 51,000 entries per sec for OC-192

  • Implementable in hardware or software.


Tcp switching can be implemented in software l.jpg

TCP Switching can be implemented in software

TCP Switching boundary router:

  • Kernel module in Linux 2.4 1GHz PC

  • Forwarding latency

    • Forward one packet: 21ms.

    • Compare to: 17ms for IP.

    • Compare to: 95ms for IP + QoS.

  • Time to create new circuit: 57ms.

Source: Byung-Gon Chun ‘01


Bandwidth inefficiencies l.jpg

Slow

Start

Congestion

Avoidance

Instantaneous

bandwidth

Circuit

BW

time

Flow

duration

Inactivity

timeout

Bandwidth inefficiencies

Compromise: inactivity timeout of few seconds


Related work l.jpg

Related work

  • IP Switching

    • Uses ATM virtual circuits (i.e. packets)

    • Became MPLS (but no longer user flows)

  • Generalized Multi-Protocol Label Switching (GMPLS)

    • Coarse circuits

    • Heavy weight signaling


Contributions43 l.jpg

Contributions

The Internet needs circuit switching

in the core

TCP Switching: how to integrate circuit switching in the core

Provisioning of fat pipes in an all-optical backbone


New networking scenario l.jpg

New networking scenario

IP routers

Optical Switches


New networking scenario45 l.jpg

New networking scenario

IP Linecards

Optical Crossconnect


Circuit creation is slow l.jpg

Circuit creation is slow

  • We need a safeguard to avoid running out of BW => inefficiency

A slow signaling requires a larger BW safeguard


Controlling coarse circuits with user flows l.jpg

Controlling coarse circuits with user flows

  • Should use the fastest optical switching elements

  • Should avoid ACKs => no RTT


Conclusions l.jpg

Conclusions

  • Circuits should be used in the core, packets in the edges

  • TCP Switching integrates circuits and packets in an evolutionary way

  • User flows can be used to control an all-optical network


Papers l.jpg

Papers

  • PMF, Nick McKeown, "TCP Switching: Exposing Circuits to IP“, IEEE Micro, 2002

  • PMF, NM, "TCP Switching: Exposing Circuits to IP“, Hot Interconnects, 2001

  • PMF, NM, “Study of routing behavior trough traffic analysis and traceroute measurements”, NAT Times, 2001

  • PMF, NM, Hui Zhang, "Is IP going to take over the world (of communications)?“, submitted


Thank you l.jpg

Thank you


  • Login