The ATM Cell Processor (MC92501) is an Asynchronous Transfer Mode (ATM)
layer device composed of dedicated high-performance ingress and egress cell
processors combined with UTOPIA Level 2-compliant physical (PHY) and switch
interface ports (see Block Diagram). The MC92501 is a second generation ATM
cell processor in MotorolaÕs 92500 series. This document provides information
on the new features offered by the second generation ATM cell processor. This
document, combined with MC92500/D, provides the complete speciÞcation for
the ATM cell processor.
New Features of the MC92501:
¥ Implements ATM Layer Functions for Broadband ISDN According to ATM
Forum UNI 4.0 and TM 4.0 SpeciÞcations, ITU Recommendations, and
Bellcore Recommendations
¥ Provides ABR Relative Rate Marking and EFCI Marking According to TM 4.0
¥ Selective Discard CLP = 1 (or CLP = 0+1) Flow on Selected Connections
¥ UTOPIA Level 2 PHY Interface and UTOPIA ATM Layer Interface
¥ Supports Both Partial Packet Discard (PPD) and Early Packet Discard (EPD)
¥ Change ABR RM Cell Priority
¥ Support for CLP Transparency
Order this document
by MC92501/D
MC92501
GC SUFFIX
GTBGA
CASE 1208
ORDERING INFORMATION
MC92501GCGTBGA
Existing MC92500 Features:
¥ Full-Duplex Operation at Data Rates up to 155 Mbit/sec
¥ Performs Internal VPI and VCI Address Compression for up to 64K VCs
¥ CLP-Aware Peak, Average, and Burst-Length Policing with Programmable
Tag/Drop Action Per Policer
¥ Supports up to 16 Physical Links Using Dedicated Ingress/Egress MultiPHY
Control Signals
¥ Each Physical Link Can Be ConÞgured as Either a UNI or NNI Port
¥ Supports Multicast, Multiport Address Translation
¥ Maintains Both Virtual Connection and Physical Link Counters on Both
Ingress and Egress Cell Flows
¥ Provides a Flexible 32-Bit External Memory Port for Context Management
¥ Automated AIS, RDI, CC, and Loopback Functions with Performance
Monitoring Block Test on All 64K Connections
¥ Programmable 32-Bit Microprocessor Interface Supporting Big-Endian or
Little-Endian Bus Formats
¥ Bidirectional UPC or NPC Design with up to Four Leaky Buckets Per
Connection
¥ Supports a Programmable Number of Additional Switch Overhead
Parameters Allowing Adaptation to Any Switch Routing Header Format
¥ Provides Per-Link Cell Counters in Both Directions
This document contains information on a new product. SpeciÞcations and information herein are subject to change without notice.
A typical ATM network consists of user end stations that
transmit and receive 53-byte data cells on virtual connections
(see Figure 1). Physical links and switching systems
interconnect the virtual connections. A virtual connectionÕs
path is established at the beginning of the data transfer,
maintained while the end-stations are communicating, and
torn down after the transfer is complete. This transmission
method increases the transfer speed because the
determination of the path the data will take is done only at the
beginning of the data transfer instead of when each data subblock or packet is transferred.
On a given physical link, each connection is assigned a
unique connection identiÞer. The connection identiÞer is
placed in the header of each cell by the transmitting equipment
and is used by the receiving equipment to route the cell to the
next physical link on the connection path. All cells belonging
to a speciÞc virtual connection follow the identical path from
VCs
Switch
the transmitting end station through the switching systems to
the receiving end station.
An ATM switch contains a high-speed switching fabric that
connects multiple line cards. The switching fabric connects the
input port to the output port based on the switchÕs routing table.
The line card interfaces between the physical medium and the
switching fabric by recovering incoming cells from the arriving
bit stream or converting outgoing cells into a bit stream for
transmission. An ATM swtich partitioned in this fashion can
efÞciently handle multiple physical links by independently
transferring each incoming ATM cell from its source port to its
destination port, based on the switchÕs routing table.
ATM standards divide the tasks to be performed on each
side of the switch fabric into PHY layer and ATM layer tasks.
The PHY layer tasks are dependent on the physical medium
that connects ATM switches. The ATM layer tasks operate at
the cell level and are independent of the physical medium.
Switch
VCs
END STATIONS
SWITCH
CLK REC
Switch
LINE
CARD
PHY
SwitchSwitchSwitch
SWITCHING FABRIC
MC92501
ATM LAYER FUNCTIONS
Figure 1. MC92501 in an ATM Network Application
END STATIONS
LINE
CARD
LINE CARD
MOTOROLA
MC92501
5
1.2.ATM Network Applications
The MC92501 performs the ATM layer functions in an ATM
switch such as cell processing and routing. Since the
MC92501 is an ATM layer device, it is PHY layer independent.
Figure 2 illustrates a typical ATM line card. The MC92501
uses an external memory for storing the cells that it
processes. In addition, the MC92501 offers an option to utilize
an external address compression device accessed via the
same external memory bus.
The microprocessor is used for conÞguration, control, and
status monitoring of the MC92501 and is responsible for
initializing and maintaining the external memory. The
MC92501 is the master of the external memory bus. At regular
intervals, the MC92501 allows the microprocessor to access
the external memory for updating and maintenance.
System RAM can also be located on the line card. The
MC92501 can support a DMA device to allow efÞcient data
transfer to this RAM without processor intervention.
RAM
MICRO-
PROCESSOR
DMA
DMA
The physical interface (PHY-IF) implements the physical
layer functions of the B-ISDN Protocol Reference Model. This
includes the physical medium dependent functions required
to transport ATM cells between the ATM user and the ATM
switch (UNI) or between two ATM switches (NNI). The cells
are transferred between the physical interface and the
MC92501 using the UTOPIA Level 2 standard.
The MC92501 implements B-ISDN UNI/NNI ATM layer
functions required to transfer cells to and from the switch over
virtual connections. These functions include usage
enforcement, address translation, and Operation,
Administration, and Maintenance (OAM) processing. The
MC92501 provides context management for up to 65,536
(64K) Virtual Connections (VCs). The VCs can be either Virtual
Path Connections (VPCs) or Virtual Channel Connections
(VCCs). ATM cells belonging to a particular VCC on a logical
link have the same unique Virtual Path IdentiÞer/Virtual
Channel IdentiÞer (VPI/VCI) value in the cell header. Similarly,
cells belonging to a particular VPC on the same logical link
share a unique VPI.
EXTERNAL
MEMORY
EXTERNAL
ADDRESS
COMPRESSION
MICROPROCESSOR BUS
RECOVERY
LINE CARD
CLOCK
PHY-IF
PHY-IF
PHY IF
MC92501
Figure 2. Typical MC92501 Line Card Application
EXTERNAL
MEMORY
BUS
TO SWITCH
FROM SWITCH
MC92501
6
MOTOROLA
SECTION 2. FUNCTIONAL DESCRIPTION
2.1.System Functional Description
A serial transmission link operating at up to 155.52 Mbit/sec
(PHY) is coupled to the MC92501 via a byte-based interface.
The transmission link timing is adapted to the MC92501 and
switch timing by means of internal cell buffers. A common clock
supplies both the PHY IF and MC92501.
The host microprocessor initializes and provides real-time
control information to the data-ßow chips (PHY IF and
MC92501) using slave accesses.
The MC92501 operates in conjunction with an external
connection memory, which provides one context entry for each
active connection. The entry consists of two types of context
parameters: static and dynamic. The static parameters are
loaded into the context memory when the VC is established,
and are valid for the duration of that connection. The static
parameters include trafÞc descriptors, OAM ßags, and ATM
switch parameters. The dynamic context parameters include
cell counters, UPC/NPC Þelds, and OAM parameters. The
dynamic parameters can be modiÞed while a particular
connection is being processed. The microprocessor can
access the external memory through the MC92501 to collect
trafÞc statistics and to update the OAM parameters. During
normal cell processing, the MC92501 has exclusive access
to the external memory and maintains external memory
coherency.
At user-programmable intervals, the MC92501 provides the
microprocessor with a Òmaintenance slot.Ó During this time,
cell processing is halted and control of the external memory
bus is relinquished. The break in cell processing is made
possible by the difference between the MC92501 cellprocessing rate and the line rate.
The microprocessor can use the maintenance slot for any
of the following tasks:
¥ Connection setup and tear down
¥ Statistics collection
¥ Updating OAM parameters of active connection
The microprocessor is responsible for the external memory
coherency during the maintenance interval.
2.2.MC92501 Functional Description
MC92501 General Features:
¥ Implements ATM layer functions for broadband ISDN
according to CCITT recommendations, ATM Forum
UNI 4.0 and TM 4.0 speciÞcations, and ITU and Bellcore
recommendations.
¥ Provides 155 Mbit/sec throughput capacity and is physical
layer independent.
¥ Optionally supports up to 16 physical links.
¥ Optionally conÞgured as a User Network Interface (UNI) or
Network Node Interface (NNI) on a per-link basis.
¥ Provides Available Bit RateÐRelative Rate (ABRÐRR)
marking and EFCI marking according to TM 4.0.
¥ Supports advanced discard policies such as Selective
Discard, Partial Packet Discard (PPD), Early Packet
Discard (EPD), and Limited Early Packet Discard (Limited
EPD).
¥ Operates in conjunction with an external memory (up to
16 MB) to provide context management for up to 64K virtual
connections.
¥ Provides cell counter coherency on a per-connection basis
by maintaining redundant copies of the counter tables and
dynamically switching between them.
¥ Provides per-link cell counters in both directions.
¥ Provides per-connection Usage Parameter Control (UPC)
or Network Parameter Control (NPC) using a leaky bucket
design with up to four buckets per connection.
¥ Provides support for Operation, Administration, and
Maintenance (OAM) Continuity Check function for all
and OAM performance monitoring on all connections.
¥ Interfaces with either big-endian or little-endian
microprocessors.
¥ Supports cell insertion into the cell streams using direct
access registers which may be written by the
microprocessor or by a DMA device.
¥ Supports copying cells from the cell streams using direct
access registers which may be read by the microprocessor
or by a DMA device.
¥ Supports multicast operation.
2.3.First Generation Features
The MC92501 is a second generation ATM cell processor
that enhances the MC92500 (Þrst generation) functionality.
The MC92501 is backwards-compatible and pincompatible with the MC92500.
This document describes the second generation
enhancements and is meant to supplement the MC92500
speciÞcation. The MC92500 speciÞcation can be ordered from
the Motorola Literature Center by requesting document
MC92500/D.
MOTOROLA
MC92501
7
SECTION 3. PACKET-BASED UPC DISCARD ALGORITHMS
3.1.Introduction
The MC92501 UPC function performs cell-based discard or
packet-based discard according to ATM Forum TM 4.0. It
supports packet discard on VC connections AAL5 packets (not
including OAM cells). The MC92501 also performs Partial
Packet Discard or Early Packet Discard.
The MC92501 offers four modes of UPC operation on a perconnection basis: Cell-Based UPC, Partial Packet Discard
(PPD), Early Packet Discard (EPD), and Limited Early Packet
Discard (Limited EPD). These modes are selected on a perconnection basis using the
Mode
bit in the Common Extension Parameters Table. Packet-
based UPC is enabled globally by the
Enable
bit in the ACR register.
IUOMÑIngress UPC Operation
IPCVÑIngress Features
3.1.1.AAL5 Packet Definition
A packet is deÞned as a stream of user cells belonging to
the same virtual connection that has a series of one or more
cells with the PTI[0] bit set to 0 and the last PTI[0] bit set to 1.
(See Figure 3.)
3.2.Cell-Based UPC
This is the default mode. The MC92501 discards cells on a
per-cell basis as deÞned in MC92500/D.
3.3.Partial Packet Discard (PPD)
According to the PPD algorithm, if a cell is discarded then
all subsequent cells belonging to that packet are discarded up
to but not including the last cell. Following is a detailed
explanation of the UPC function.
¥ The UPC is a two-state machine: discarding and not-
discarding. See Figure 4.
¥ While the UPC is in the not-discarding state, it performs
normal cell-based operation with tagging and policing
counter updates.
¥ The UPC transitions from the not-discarding to the
discarding state on the Þrst discarded cell.
¥ While the UPC is in the discarding state, it does not update
the UPC bucket but it does increment the policing discard
counter.
¥ When in the discarding state and the last cell of a packet
is received, there are two options:
Ñ If
all the cells belonging to that packet were discarded,
then this last cell is discarded.
Ñ If
not all the cells belonging to that packet were
discarded, then this means that the packet was
truncated and this last cell is admitted in order to
delineate the corrupted packet from the next packet.
There is however one exception: if this last cell is
violating cell-based UPC then it is discarded.
Figure 5 illustrates an example for the PPD algorithm. A
UPC policy violation occurs during the transmission of the Þrst
packet. The UPC detects the violation and discards the
remainder of the packet except for the last cell. The last cell
of the Þrst packet is transmitted to avoid the concatenation of
the corrupted packet with the subsequent Packet #2. If the
UPC detects that the Þrst cell of Packet #3 violates its policy
then Packet #3 is truncated. Packet #3Õs last cell is not
transmitted because it cannot be admitted by the cell-based
UPC. Packet #4 is not transmitted either because its Þrst cell
violates the UPC policy.
Figure 3. Delineation of a Packet Within a Cell Stream
LAST CELL OF PACKET ARRIVES
If all cells within a packet are discard
OR
If the last cell is violating UPC policy
then discard the last cell
If not all calls within packet discarded, then admit last cell
Figure 4. UPC Discarding State Machine
PACKET 1
NOT-DISCARDING
DISCARDING
FIRST CELL OF
PACKET 2
CELL VIOLATES
UPC POLICY
MC92501
8
MOTOROLA
Input Stream
UPC Discard Decision
Packet #1Packet #2Packet #3 Packet #4
Output Stream
L
Figure 5. Partial Packet Discard
3.4.Early Packet Discard (EPD)
According to the EPD algorithm, the decision to discard a
packet takes place only at the beginning of a packet. This
means that the complete packet is either fully discarded or
fully passed. The following explains how EPD is implemented.
¥ When the EPD is discarding cells, the buckets are not
updated but the policing discard counter is incremented.
¥ When the EPD decides that a frame should be passed this
means that:
Ñ All tagging buckets continue to work in a cell-based
fashion.
Ñ All discarding buckets perform their calculations as if
the limit parameter is inÞnite, and therefore increment
the bucket content and do not discard any cells. As a
result, their bucket content can be greater than their
bucket limit.
Ñ The MC92501 may increment its police tagging counter.
Figure 6 illustrates an example for the EPD algorithm. A cell
within the Þrst packet violates the UPC, but due to EPD this
packet is fully passed. Since the Þrst cell of Packet #2 violates
the UPC, the second packet is fully discarded. Likewise, cells
within Packet #3 violate the UPC, but this packet is not
discarded. Since the fourth packet comes after a relatively
long time, which allows the UPC buckets to drain, Packet #4Õs
cells do not violate the UPC policy.
3.5.Limited Early Packet Discard (Limited
EPD)
One disadvantage of the EPD algorithm is that once it
decides to admit a packet it cannot change its decision until
the last cell of that packet. In the case of big packets, the switch
can run into congestion. Using the Limited EPD algorithm, a
connection can stop passing cells because of EPD once it
reaches a predeÞned limit. That limit, in the case of the
MC92501, is reached once the
cells. The Þrst bucket should have the same parameters as
one of the other buckets except for the limit, which is bigger.
Figure 7 describes a UPC which contains three buckets.
The Þrst bucket is for limiting EPD, and there are two other
buckets. The Þrst and second buckets share the same
parameters except for the limit. Therefore, their bucket content
is always the same, although the second bucketÕs content is
higher than its limit and cells are admitted by the EPD
algorithm. When the Þrst bucket reaches its limit, then cells
will be discarded.
Figure 8 describes the EPD and Limited EPD functions.
Þrst bucket starts discarding
Input Stream
UPC Discard Decision
Output Stream
MOTOROLA
Packet #1Packet #2
Figure 6. Early Packet Discard
Packet #3
Packet #4
MC92501
9
Limit
Limit
Limit
First BucketSecond BucketThird Bucket
Figure 7. Limited Early Packet Discard
.
Input Stream
First Bucket Limit
Other BucketsÕ Limit
EPDÐUPC Output Stream
Limited EPDÐUPC Output Stream
Figure 8. Difference Between Early Packet Discard and Limited Early Packet Discard
Packet #1Packet #2
Packet #3
Packet #4
MC92501
10
MOTOROLA
SECTION 4. SELECTIVE DISCARD
ATM Forum TM 4.0 deÞnes procedures according to which
cells can be discarded by network elements. A switching
element may discard cells belonging to selected connections
or cells whose CLP = 1 in case of congestion. This function is
called selective discard and it is implemented by the
MC92501. Selective discard is enabled by the
Ingress Congestion NotiÞcation
Control Register (IPLR). Selective discard can be enabled on
bit in the Ingress Processing
ICNGÑGlobal
a per-connection basis by the
Discard Operation Mode
Extension Word. This Þeld determines whether selective
discard is enabled and whether selective discard is performed
on CLP = 1 or on CLP = 0+1 trafÞc. Selective discard can be
enabled globally by the
CFB ConÞguration Register (ACR).
Þeld in the Common Parameters
IPCVÑIngress Enable bit in the ATMC
ISDMÑIngress Selective
MOTOROLA
MC92501
11
SECTION 5. AVAILABLE BIT RATE (ABR) SUPPORT
5.1.Overview and Features
The MC92501 provides a full Available Bit Rate (ABR)
solution for switch behavior relative rate marking and EFCI
marking in accordance with ATM Forum TM 4.0. It also
provides the switch fabric with an interface to increase the RM
cellsÕ trafÞc priority. Following is a list of features:
¥ Performs Relative Rate (RR) marking on Forward
Resource Management (FRM) and/or Backward Resource
Managment (BRM) cells, on selected connections. This
feature is enabled by either setting the ATMC CFB
ConÞguration RegisterÕs (ACR) VP RM Cell PTI (NPRP) bit
or by setting the PTI Þeld in the cellÕs header to Ò110BÓ.
¥ Performs EFCI marking on non-RM cells whose PTI[2] =
0, on selected connections. This feature is enabled by
either control registers or by Þelds that it gets from the
overhead of cells which are received from the switch fabric.
¥ Resets EFCI on non-RM cells whose PTI[2] = 0, on
selected connections.
Header = 5 bytesPayload = 48 bytes
GFC/VPI
VPI
VCI
PTICLPHEC
8
PIDERCCRMCRQLSN
2 x 84 x 84 x 82 x 82 x 8
830 x 8 + 6
¥ Checks CRC on received RM cells and generates CRC for
transmitted RM cells.
¥ Provides different priority to RM cells.
¥ Can copy RM cells to the microprocessor or remove them
from the ßow.
5.2.RM Cell Definition
A cell is an RM cell if and only if at least one of the following
conditions is met:
¥ The cell belongs to a VC connection and its PTI = 6.
¥ The cell belongs to a VP connection, its VCI = 6, and its
PTI = 6.
¥ The cell belongs to a VP connection, its VCI = 6, and the
ATMC CFB ConÞguration Register has the
Cell PTI
bit set.
VPRPÑVP RM
5.3.RM Cell Fields
10
Reserved
CRC-10
DIR
BN
1
1
NOTES:
PID= 1
DIR= Direction
¥ 0 = Forward RM cell
¥ 1 = Backward RM cell
BN= Backward Explicit Congestion
¥ 0 = Generated by source
¥ 1 = Generated not by the source
CI= Congestion Indication
NI= No Increase Bit
ER= Explicit Rate
CCR = Current Cell Rate
MCR = Minimum Cell Rate
CRC - 10
CI 1NI 1RA
1
Figure 9. RM Cell Fields
Reserved
3
MC92501
12
MOTOROLA
5.4.Cell Marking (CI, NI, PTI)
Figure 10 illustrates two MC92501 devices connected to a
switch fabric. In this example, the ABR ßow travels from left to
right. This means that data cells are ßowing from left to right,
FRM cells are ßowing from left to right, and BRM cells are
ßowing from right to left. The switch marks FRM and user cells
Ingress Flow Status
Ingress User Cell Marking (EFCI)
Ingress FRM Cell Marking (CI or NI)
Egress BRM Cell Marking (CI or NI)
EFCI, FRM
BRM
Ingress
Switch Fabric
Egress
MC92501 #1MC92501 #2
ßowing downstream, and BRM cells ßowing upstream. This
switch function can be implemented in the ingress of MC92501
#1 and in the egress of MC92501 #2. MC92501 #1 marks cells
because of the ingress ßow status (for example, ingress ßow
congestion) while MC92501 #2 marks cells because of the
egress ßow status.
Egress Flow Status
Egress User Cell Marking (EFCI)
Egress FRM Cell Marking (CI or NI)
Ingress BRM Cell Marking (CI or NI)
Egress
Ingress
Downstream Direction
Upstream Direction
Figure 10. ABR Flow Cell Marking Example
The MC92501 can take the following actions in response to
the ingress ßow status:
¥ Perform EFCI marking on ingress cells; i.e., set PTI[1] bit
in cells on which PTI[2] = 0.
¥ Set CI or NI in ingress FRM cells.
¥ Set CI or NI in egress BRM cells.
The MC92501 can take the following actions in response to
the egress ßow status:
¥ Perform EFCI marking on egress cells; i.e., set PTI[1] bit
in cells on which PTI[2] = 0.
¥ Set CI or NI in egress FRM cells.
¥ Set CI or NI in ingress BRM cells.
Figure 11 is an overview of the MC92501 marking scheme.
MOTOROLA
MC92501
13
OR
Global Reg.
CellÕs Overhead
Context Bit
Ingress
Status
Collection
Global Registers
Ingress Flow Status
Ingress
Action:
Marking
Global Registers
Context Bits
Cell Type
Set CI
Set NI
Set PTI
Global Reg.
CellÕs Overhead
Context Bit
Egress
Status
Collection
Egress Flow Status
Figure 11. Cell Marking Scheme
There are various ways to inform the MC92501 that it should
mark a cell due to the ingress ßow status or the egress ßow
status. This scheme also shows that the status of the ingress
ßow, the status of the egress ßow, global registers, a context
bit, and the cell type impact the decision of setting CI, NI, and
PTI. Following is a detailed description of each of the function
boxes.
5.4.1.Sources for Ingress Flow Status
The ingress ßow status is gathered from three sources:
global register, cellÕs overhead, or context bit.
5.4.1.1.Ingress Flow Status from Global Register
The switch fabric can notify the MC92501 that it should mark
cells because of the ingress ßow status by setting the
Global Ingress ABR Mark Enable
bit in the Ingress Processing
Control Register (IPLR).
5.4.1.2.Ingress Flow Status from CellÕs Overhead
The switch fabric can notify the MC92501 that it should mark
cells because of the ingress ßow status of connection #n by
setting the
IFSÑOverhead Ingress Flow Status bit in the
overhead of egress cells belonging to that connection. The
location of this bit in the overhead is programmable using the
EIBYÑIFS Byte Location bit and the EIBIÑIFS Bit Location
bit in the Egress Switch Overhead Information Register 1
(ESOIR1). This bit is enabled by the
EIASÑGlobal IFS Enable
bit in the Egress Switch Interface ConÞguration Register
(ESWCR). The MC92501 can be programmed that in such a
case it will mark egress BRM cells.
IAMEÑ
Set CI
Egress
Action:
Marking
Section 5.4.1.2 for details on enabling of
Ingress Flow Status
receives that cell, it copies the bit into the
Ingress Flow Status
Set NI
Set PTI
IFSÑOverhead
bit and its location.) When the MC92501
CIFSÑConnection
bit in the Common Parameters Extension
Word of connection #n. The MC92501 can be programmed
that in such a case it will mark ingress FRM cells or perform
EFCI marking.
5.4.1.4.Logic of Ingress Flow Status
The ingress ßow status equals 1 if:
IAME = 1
IFS = 1 and EIAS = 1 and egress = 1 OR
CIFS = 1 and EIAS = 1 and ingress = 1
Where:
IAME = Global Ingress ABR Mark Enable
IFS = Overhead Ingress Flow Status
EIAS = Global IFS Enable
CIFS = Connection IFS Enable
Egress = Programmed Overhead Egress Bit
Ingress = Programmed Overhead Ingress Bit
5.4.2.Sources for Egress Flow Status
The egress ßow status is gathered from three sources:
global register, cellÕs overhead, and context memory.
5.4.2.1.Egress Flow Status from Global Register
The switch fabric can notify the MC92501 that it should mark
cells because of the egress ßow status by setting the
Global Egress ABR Mark Enable
bit in the Egress Processing
EAMEÑ
Control Register (EPLR).
5.4.1.3.Ingress Flow Status from Context Memory
The switch fabric can notify the MC92501 that it should mark
cells because of the ingress ßow status of connection #n by
setting the
IFSÑOverhead Ingress Flow Status bit in the
overhead of egress cells belonging to that connection. (See
MC92501
14
5.4.2.2.Egress Flow Status from CellÕs Overhead
The switch fabric can notify the MC92501 that it should mark
cells because of the egress ßow status of connection #n by
setting the
EFSÑOverhead Egress Flow Status bit in the
overhead of egress cells belonging to that connection. The
MOTOROLA
location of this bit in the overhead is programmable using the
EEBYÑEFS Byte Location bit and the EEBIÑEFS Bit
Location
Register 1 (ESOIR1). This bit is enabled by the
EFS Enable
Register (ESWCR). The MC92501 can be programmed that
in such a case it will mark egress FRM cells or perform EFCI
marking.
5.4.2.3.Egress Flow Status from Context Memory
cells because of the egress ßow status of connection #n by
setting the
overhead of egress cells belonging to that connection. (See
Section 5.4.2.2 for details on enabling of EFSÑOverheadEgress Flow Status bit and its location.) When the MC92501
receives that cell, it copies the bit into the CEFSÑConnectionEgress Flow Status bit in the Common Parameters Extension
Word of connection #n. The MC92501 can be programmed
that in such a case it will mark ingress BRM cells.
5.4.2.4.Logic of Egress Flow Status
5.4.3.Ingress ABR Marking Bits
ßow status or egress ßow status.
MC92501 can perform one or more of the following:
¥ Set CI bit in an ingress FRM cell Ñ when the ISFCEÑ
¥ Set NI bit in an ingress FRM cell Ñ when the ISFNEÑ
¥ Set PTI[1] bit in an ingress cell whose PTI[2] = 0 Ñ when
MC92501 can perform one or more the following:
¥ Set CI bit in an ingress BRM cell Ñ when the ISBCEÑ
¥ Set NI bit in an ingress BRM cell Ñ when the ISBNEÑ
bit in the Egress Switch Overhead Information
EEASÑGlobal
bit in the Egress Switch Interface ConÞguration
The switch fabric can notify the MC92501 that it should mark
EFSÑOverhead Egress Flow Status bit in the
The egress ßow status equals 1 if:
EAME = 1 OR
EFS = 1 and EEAS = 1 and egress = 1 OR
CEFS = 1 and EEAS = 1 and ingress = 1
Where:
EAME = Global Egress ABR Mark Enable
EFS = Overhead Egress Flow Status
EEAS = Global EFS Enable
CEFS = Connection EFS Enable
Egress = Programmed Overhead Egress Bit
Ingress = Programmed Overhead Ingress Bit
The MC92501 can mark cells as a result of either ingress
In the case where ingress ßow status is asserted, the
Global Ingress Set FRM CI Enable bit in the Ingress
Processing ConÞguration Register (IPCR) is set.
Global Ingress Set FRM NI Enable bit in the IPCR is set.
the ISPEÑGlobal Ingress Set PTI Enable bit in the IPCR
is set.
In the case where egress ßow status is asserted, the
Global Ingress Set BRM CI Enable bit in the IPCR is set.
Global Ingress Set BRM NI Enable bit in the IPCR is set.
All cell marking on the ingress is enabled on a per-
connection basis by the CIMEÑConnection Ingress Marking
Enable bit in the Common Parameters Extension Word.
5.4.3.1.Logic of Ingress ABR Marking Bits
The CI bit is set if:
FRM cell and CIME = 1 and ingress ßow status = 1
and ISFCE = 1 OR
BRM cell and CIME = 1 and egress ßow status = 1
and ISBCE = 1
The NI bit is set if:
FRM cell and CIME = 1 and ingress ßow status = 1 and
ISFNE = 1 OR
BRM cell and CIME = = 1 and egress ßow status = 1 and
ISBNE = 1
The PTI[1] bit is set if:
PTI[2] = 0 and CIME = 1 and ingress ßow status = 1 and
ISPE = 1
Where:
CIME = Connections Ingress Marking Enable
FRM Cell = Cell marked as FRM cell
BRM Cell = Cell marked as BRM cell
Ingress Flow Status = Set as deÞned in Section 5.4.1.4
Egress Flow Status = Set as deÞned in Section 5.4.2.4
ISFCE = Global Ingress Set FRM CI Enable
ISFNE = Global Ingress Set FRM NI Enable
ISPE = Global Ingress Set PTI Enable
ISBCE = Global Ingress Set BRM CI Enable
ISBNE = Global Ingress Set BRM NI Enable
5.4.4.Egress ABR Marking Bits
The MC92501 can mark cells as a result of either ingress
ßow status or egress ßow status.
In the case where egress ßow status is asserted, the
MC92501 can perform one or more of the following:
¥ Set CI bit in an egress FRM cell Ñ when the ESFCEÑ
Global Egress Set FRM CI Enable bit in the Egress
Processing ConÞguration Register (EPCR) is set.
¥ Set NI bit in an egress FRM cell Ñ when the ESFNEÑ
Global Egress Set FRM NI Enable bit in the EPCR is set.
¥ Set PTI[1] bit in an egress cell whose PTI[2] = 0 Ñ when
the ESPEÑGlobal Egress Set PTI Enable bit in the EPCR
is set.
In the case where ingress ßow status is asserted, the
MC92501 can perform one or more the following:
¥ Set CI bit in an egress BRM cell Ñ when the ESBCEÑ
Global Egress Set BRM CI Enable bit in the EPCR is set.
¥ Set NI bit in an egress BRM cell Ñ when the ESBNEÑ
Global Egress Set BRM NI Enable bit in the EPCR is set.
All cell marking on the egress is enabled on a per-
connection basis by the CEMEÑConnection Egress MarkingEnable bit in the Common Parameters Extension Word.
MOTOROLA
MC92501
15
5.4.4.1.Logic of Egress ABR Marking Bits
The CI bit is set if:
FRM cell and CEME = 1 and egress ßow status = 1
and ESFCE = 1 OR
BRM cell and CEME = 1 and ingress ßow status = 1
and ESBCE = 1
The NI bit is set if:
FRM cell and CEME = 1 and egress ßow status = 1 and
ESFNE = 1 OR
BRM cell and CEME = 1 and ingress ßow status = 1 and
ESBNE = 1
The PTI[1] bit is set if:
PTI[2] = 0 and CEME = 1 and egress ßow status = 1 and
ESPE = 1
Where:
CEME = Connections Egress Marking Enable
FRM Cell = Cell marked as FRM cell
BRM Cell = Cell marked as BRM cell
Ingress Flow Status = Set as deÞned in Section 5.4.1.4
Egress Flow Status = Set as deÞned in Section 5.4.2.4
ESFCE = Global Egress Set FRM CI Enable
ESFNE = Global Egress Set FRM NI Enable
ESPE = Global Egress Set PTI Enable
ESBCE = Global Egress Set BRM CI Enable
ESBNE = Global Egress Set BRM NI Enable
5.4.5.Cell Marking Examples
Figure 12, Figure 13, and Figure 14 provide examples for
CI and NI marking.
NOTE 1
NOTE 3
NOTE 4
MC92501 #1MC92501 #2
NOTES:
1. Initially the microprocessor conÞgures the MC92501 #1 as follows:
¥ Sets the ISFCEÑGlobal Ingress Set FRM CI Enable bit.
¥ Sets the CIMEÑConnection Ingress Marking Enable bit for selected ABR connections.
2. The switch fabric informs the microprocessor that ingress ABR queues have reached
some limit.
3. The microprocessor sets the IAMEÑGlobal Ingress ABR Mark Enable bit.
4. The MC92501 sets the CI bit for FRM cells belonging to the selected ABR connections.
Microprocessor
NOTE 2
Switch Fabric
Downstream Direction
Upstream Direction
Figure 12. Enable Marking CI Bits of Ingress FRM Cells
MC92501
16
MOTOROLA
Loading...
+ 36 hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.