review and update of carrier extension...

of 22

11-November-1996

al

IEEE P802.3zGigabit Task Force

Howard M. Frazier, Jr.

Review and Update ofCarrier Extension Propos

Howard M. Frazier, Jr.Sun Microsystems Computer Company

Internet and Networking Products Group11-November-1996

IEEE 802.3z TF

of 22

11-November-1996



Outline

■ Introduction

■ Topologies

■ Bit Budget

■ Issues

■ Solution

■ Impact on MAC

■ MAC Parameters

■ Impact on Reconciliation Sublayer

■ Performance

■ Conclusions

of 22

11-November-1996

/s switched

mplicated than

/s

ulously

sential



Introduction

■ Scaling CSMA/CD to 1000 Mb/s is a good idea

■ Widely implemented, well understood protocol

■ Demonstrated low cost at 10 and 100 Mb/s

■ Cost benefit of 1000 Mb/s shared versus 1000 Mb

■ Scaling CSMA/CD to 1000 Mb/s is slightly more coscaling it to 100 Mb/s

■ Wire delays are 10x larger (in BT) than at 100 Mb

■ “shift the decimal point” approach results in a ridicsmall collision domain diameter

■ Backwards compatibility with 10 and 100 Mb/s is es

of 22

11-November-1996

repeater

domain

d desktops

desktops



Topologies

■ Most useful CSMA/CD topologies have at least one

CSMA/CD used in a single repeater collisionwithin a server room

1000 Mb/s repeater

100/1000 switches

10/1000 switches

high en

low end

servers

1000

1000

1000

1000

1000

100

10

of 22

11-November-1996

omains

esktops

sktops

d desktops



Topologies (cont)

CSMA/CD used in single repeater collision dwith 1000 Mb/s horizontal runs

1000 Mb/s switch

1000 Mb/s repeaters

100/1000 switches

higher end d

high end de

servers &ultra high en

1000

1000

1000

1000

1000

1000

100

of 22

11-November-1996

sktops must



Topologies (cont)

■ Horizontal runs are 100 meters maximum

■ A CSMA/CD topology with horizontal runs out to desupport a diameter of at least 200 meters

100 m100 m

200 m

of 22

11-November-1996

d bit budget



Bit Budgets

■ Bit budget calculations are highly dependent on:

■ Physical layer signalling method and architecture

■ MAC <=> PHY data path width

■ MAC state machine frequency

■ Repeater data path width

■ Repeater state machine frequency

■ Fairness issues

■ See Stephen Haddock’s presentation for an updateanalysis

of 22

11-November-1996

,

tic

12 bits

es



Bit Budgets (cont)

■ Conclusions from bit budget analysis

■ Cable delay dominates, but even with short cablesthe bit budget exceeds 512 BT

■ DTE and repeater delay estimates may be optimis

■ The minimum frame size must be increased from 5to achieve useful topologies at 1000 Mb/s

■ Recommend a new minimum frame size of 512 Byt

■ Simply change bits to bytes!

of 22

11-November-1996

ameSize

x linksm size

10 Mb/s networks,

e 802.3 MAC when format

only onher Ethernet or



Issues

■ It’s not that simple

■ Trying to solve the problem by increasing the minFrto 512 Bytes has harmful side effects

■ Degrades maximum small packet rate on full duplewhere it is not necessary to use the larger minimu

■ Can’t propagate the inflated packets to 100 Mb/s orthe performance loss would be hideous

■ Can’t rely on the padding/stripping mechanism in thusing protocol stacks which use the Ethernet frame

■ Must find a way to extend the minimum frame size 1000 Mb/s, CSMA/CD networks, regardless of whet802.3 frame format is used

of 22

11-November-1996

nt without

ls

ation

d passing the



Solution

■ Conceptually, extend the duration of the carrier eveextending the data field, or altering the FCS field

■ Introduce a new mechanism which appendsnon-data symbols to the end of short frames

■ Extend the collision window to include these symbo

■ Include the extension in the fragment discard calcul

■ Remove the extension before checking the FCS anframe to LLC

of 22

11-November-1996

are recognized

n integer

FCS Extension



Solution (cont)

■ Frame extension

■ The extension symbols are non-data symbols, and as such by the MAC and the PHY

■ The extension starts on an octet boundary, and is anumber of octets in length

Preamble SFD DA SAType/Length

Data

64 bytes min

512 bytes min

FCS Coverage

Duration of Carrier Event

of 22

11-November-1996

e 4.4}

end of frame}



Impact on MAC

■ New const (4.2.7.1)const

extendSize = ...; {in bits, implementation-dependent, seextensionBit = ...; {a new type of non-data bit}

■ New Transmit State Variable (4.2.7.2)var

extension:0..extendSize; {length of extension}

■ New Receive State Variable (4.2.7.3)var

extendCount: 0..extendSize; {count of extension bits at

of 22

11-November-1996



Impact on MAC

■ Mod to StartTransmit (4.2.8)procedure StartTransmit;begin

extension := 0;currentTransmitBit := 1;lastTransmitBit := frameSize;transmitSucceeding := true;transmitting := true;lastHeaderBit := headerSize

end; {StartTransmit}

■ Mod to StartReceive (4.2.9)procedure StartReceive;

begincurrentReceiveBit :=1;extendCount := 0;receiving := true

end; {StartReceive}

of 22

11-November-1996

+ extendSize);

itBit]);



Impact on MAC (Transmitter)

■ Mod to BitTransmitter (4.2.8)process BitTransmitter;begin

cycle {outer loop}if transmitting thenbegin {inner loop}

if halfDuplex then (extension : = minFrameSizePhysicalSignalEncap; {send preamble and sfd}while transmitting dobegin

if currentTransmitBit > lastTransmitBit thenTransmitBit(extensionBit)

elseTransmitBit(outgoingFrame[currentTransm

if newCollision then StartJam else NextBitend;

end; {inner loop}end; {outer loop}

end; {BitTransmitter}

of 22

11-November-1996

it)




■ Mod to NextBit (4.2.8)procedure NextBit;

begincurrentTransmitBit := currentTransmitBit + 1;transmitting := ((currentTransmitBit ≤ lastTransmitB

or (currentTransmitBit ≤ extension))end; {NextBit}

■ Mod to StartJam (4.2.8)procedure StartJam;

begincurrentTransmitBit := 1;lastTransmitBit := jamSize;extension := 0;newCollision := false

end; {StartJam}

of 22

11-November-1996

extendSize) then




■ Mod to WatchForCollision (4.2.8)procedure WatchForCollision;begin

if transmitSucceeding and collisionDetect thenbegin

if currentTransmitBit > (minFrameSize - headerSize +lateCollisionError := true;

newCollision:= true;transmitSucceeding := false

endend {WatchForCollision}

of 22

11-November-1996

sfd}a Access}



Impact on MAC (Receiver)

■ Mod to BitReceiver (4.2.9)process BitReceiver;

var b: Bit;begin

cycle {outer loop}while receiving dobegin {inner loop}if currentReceiveBit = 1 then

PhysicalSignalDecap; {strip off the preamble andb := ReceiveBit; {get next bit from physical Mediif receiveDataValid then

if b = extensionBit thenextendCount := extendCount + 1elsebegin {append bit to frame}

incomingFrame[currentReceiveBit] := b;currentReceiveBit := currentReceiveBit + 1

endreceiving := receiveDataValidend {inner loop}frameSize := currentReceiveBit - 1 + extendCount

end {outerloop}end; {BitReceiver}

of 22

11-November-1996

rriving}

t boundary}

ize + extendSize));

);



Impact on MAC (Receiver)

■ Mod to ReceiveLinkMgmt (4.2.9)procedure ReceiveLinkMgmt;begin

repeatStartReceive;while receiving do nothing; {wait for frame to finish aexcessBits := frameSize mod 8;frameSize := frameSize - excessBits; {truncate to octeif halfDuplex thenbegin

receiveSucceeding :=((frameSize ≥ ³(minFrameSframeSize := frameSize - extendCount

endelse

receiveSucceeding := (frameSize ≥ minFrameSizeuntil receiveSucceeding

end;

of 22

11-November-1996



MAC Parameters

■ New Parameter Table 4.4.2.4

Parameters Values

slotTime 4096 bit timesinterFrameGap 96 nsattemptLimit 16backoffLimit 10jamSize 32 bitsmaxFrameSize 1518 octetsminFrameSize 512 bits (64 octets)addressSize 48 bitsextendSize 448 octets

of 22

11-November-1996

ATA.request

d by the

ATA.indicate

d by the



Impact on Reconciliation Sublayer

■ 22.2.1.1.2 Semantics of the service primitive PLS_D

■ Add EXTEND to the range of values communicateOUTPUT_UNIT parameter

■ 22.2.1.2.2 Semantics of the service primitive PLS_D

■ Add EXTEND to the range of values communicateINPUT_UNIT parameter

of 22

11-November-1996

P

tribution

sults using



erformance

■ Simulations used “workgroup average” packet size disderived from some “real world” sampling

■ See Mohan Kalkunte’s presentations for simulation re“workgroup average” packet size distribution

of 22

11-November-1996

/s

ate



Conclusions

■ Useful CSMA/CD networks can be built at 1000 Mb

■ The performance of the baseline proposal is adequ

review and update of carrier extension...

Documents