Network Working Group K. Murakami
Request for Comments: 2174 M. Maruyama
Category: Informational NTT Laboratories
June 1997
A MAPOS version 1 Extension - Switch-Switch Protocol
Status of this Memo
This memo provides information for the Internet community. This memo
does not specify an Internet standard of any kind. Distribution of
this memo is unlimited.
Authors' Note
This memo documents a MAPOS (Multiple Access Protocol over SONET/SDH)
version 1 extension, Switch Switch Protocol which provides dynamic
routing for unicast, broadcast, and multicast. This document is NOT
the product of an IETF working group nor is it a standards track
document. It has not necessarily benefited from the widespread and
in depth community review that standards track documents receive.
Abstract
This document describes a MAPOS version 1 extension, SSP (Switch
Switch Protocol). MAPOS is a multiple access protocol for
transmission of network-protocol packets, encapsulated in High-Level
Data Link Control (HDLC) frames, over SONET/SDH. In MAPOS network, a
SONET switch provides the multiple access capability to end nodes.
SSP is a protocol of Distance Vector family and provides unicast and
broadcast/multicast routing for multiple SONET switch environment.
1. Introduction
This document describes an extension to MAPOS version 1, Switch
Switch Protocol, for routing both unicast and broadcast/multicast
frames. MAPOS[1], Multiple Access Protocol over SONET (Synchronous
Optical Network) / SDH (Synchronous Digital Hierarchy) [2][3][4][5],
is a link layer protocol for transmission of HDLC frames over
SONET/SDH. A SONET switch provides the multiple access capability to
each node. SSP is a dynamic routing protocol designed for an
environment where a MAPOS network segment spans over multiple
switches. It is a protocol of Distance Vector family. It provides
both unicast and broadcast/multicast routing. First, this document
describes the outline of SSP. Next, it explains unicast and
broadcast/multicast routing algorithms. Then, it describes the SSP
protocol in detail.
Murakami & Maruyama Informational [Page 1]RFC 2174 MAPOS June 1997
2. Constraints in Designing SSP
SSP is a unified routing protocol supporting both unicast and
broadcast/multicast. The former and the latter are based on the
Distance Vector [6][7] and the spanning tree[8] algorithm,
respectively. In MAPOS version 1, a small number of switches is
assumed in a segment. Thus, unlike DVMRP(Distance Vector Multicast
Routing Protocol)[8], TRPB(Truncated Reverse Path Broadcasting) is
not supported for simplicity. This means that multicast frames are
treated just the same as broadcast frames and are delivered to every
node.
In MAPOS version 1, there are two constraints regarding design of the
broadcast/multicast routing algorithm;
(1) there is no source address field in MAPOS HDLC frames
(2) there is no TTL(Time To Live) field in MAPOS HDLC frames to
prevent forwarding loop.
To cope with the first issue, VRPB(Virtual Reverse Path Broadcast)
algorithm is introduced. In VRPB, all broadcast and multicast frames
are assumed to be generated by a node under a specific switch called
VSS(Virtual Source Switch). VSS is the switch which has the smallest
switch number in a MAPOS network. Each switch determine its place in
the spanning tree rooted from VSS independently. Whenever a switch
receives a broadcast/multicast frame, it forwards the frame to all
upstream and downstream switches except for the one which has sent
the frame to the local switch.
To cope with the second issue, the forward delay timer is introduced.
Even if a switch finds a new VSS, it suspends forwarding for a time
period. This timer ensures that all the switches have a consistent
routing information and that they are synchronized after a topology
change.
3. Unicast Routing in SSP
This section describes the address structure of MAPOS version 1 and
the SSP unicast routing based on it.
Murakami & Maruyama Informational [Page 2]RFC 2174 MAPOS June 1997
3.1 Address Structure of MAPOS version 1
In a multiple switch environment, a node address consists of the
switch number and the port number to which the node is connected. As
shown in Figure 1, the address length is 8 bits and the LSB is always
1, which indicates the end of the address field. A MSB of 0 indicates
a unicast address. The switch and the port number fields are
variable-length. In this document, a unicast address is represented
as "0 <switch-number> <port number>". Note that a port number
includes EA bit.
MSB of 1 indicates multicast or broadcast. In the case of broadcast,
the address field contains all 1s (0xff in hex). In the case of
multicast, the remaining bits indicate a group address. The switch
number field is variable-length. A multicast address is represented
as "1 <group address>".
Switch Number(variable length)
|
| +--- Port Number
| |
V V
|<->|<------->|
+-------------+-+
| | | | | | | | |
| | |1|
+-+-----------+-+
^ ^
| |
| +------- EA bit (always 1)
|
1 : broadcast, multicast
0 : unicast
Figure 1 Address Format
Figure 2 shows an example of a SONET LAN that consists of three
switches. In this configuration, two bits of a node address are used
to indicate the switch number. Node N1 is connected to port
0x03(000011 in binary) of the switch S2 numbered 0x2. Thus, the node
address is 01000011 in binary. Node N4 has an address 01101001 in
binary since the connected switch number is 0x3 and the port number
is 0x09.
Murakami & Maruyama Informational [Page 3]RFC 2174 MAPOS June 1997
01000011
+------+
| node |
| N1 |
+------+
01000101 |0x03 |0x03 00101001
+------+ +---+----+ +---+----+ +------+
| node +-----+ SONET +---------+ SONET +------+ node |
| N2 | 0x05| Switch |0x09 0x05| Switch |0x09 | N3 |
+------+ | S2 | | S1 | +------+
| (0x2) | | (0x1) |
+---+----+ +---+----+
|0x07 |0x07
| |
| |0x03 01101001
| +---+----+ +------+
+--------------+ SONET +-----+ node |
0x05| Switch |0x09 | N4 |
| S3 | +------+
| (0x3) |
+---+----+
|0x07
Figure 2 Multiple SONET Switch Environment
3.2 Forwarding Unicast Frames
Unicast frames are forwarded along the shortest path. For example, a
frame from node N4 destined to N1 is forwarded by switch S3 and S2.
These SONET switches forwards an HDLC frame based on the destination
switch number contained in the destination address.
Each switch keeps a routing table with entries for possible
destination switches. An entry contains the subnet mask, the next hop
to the adjacent switch along the shortest path to the destination,
the metric measuring the total distance to the destination, and other
parameters associated with the entry such as timers. For example, the
routing table in switch S1 will be as shown in Table 1. The metric
value 1 means that the destination switch is an adjacent switch. The
value 16 means that it is unreachable. Although the values between 17
and 31 also mean unreachable, they are special values utilized for
split horizon with poisoned reverse [8].
Murakami & Maruyama Informational [Page 4]RFC 2174 MAPOS June 1997
+-------------------------+----------+--------+------------+
| destination | subnet | next hop | metric | other |
| switch | mask | port | | parameters |
+-------------+-----------+----------+--------+------------+
| 01000000 | 11100000 | 00000101 | 1 | |
+-------------+-----------+----------+--------+------------+
| 01100000 | 11100000 | 00000111 | 1 | |
+-------------+-----------+----------+--------+------------+
Table 1 An Example of a Routing Table
When a switch receives a unicast frame, it extracts the switch number
from the destination address. If it equals to the local switch
number, the frame is sent to the local node through the port
specified in the destination address. Otherwise, the switch looks up
its routing table for a matching destination switch number by masking
the destination address with the corresponding subnet mask. If a
matching entry is found, the frame is sent to an adjacent switch
through the next hop port in the entry. Otherwise, it is silently
discarded or sent to the control processor for its error processing.
3.4 Protocol Overview
This subsection describes an overview of the unicast routing protocol
and its algorithm.
3.4.1 Route Exchange
SSP is a distance vector protocol to establish and maintain the
routing table. In SSP, each switch sends a routing update message to
every adjacent switches every FULL_UPDATE_TIME (10 seconds by
default). The update message is a copy of the routing table, that is,
routes.
When a switch receives an update message from an adjacent switch
through a port, it adds the cost associated with the port, usually 1,
to every metric value in the message. The result is a set of new
metrics from the receiving switch to the destination switches. Next,
it compares the new metrics with those of the corresponding entries
in the existing routing table. A smaller metric means a better route.
Thus, if the new metric is smaller than the existing one, the entry
is updated with the new metric and next hop. The next hop is the port
from which the update message was received. Otherwise, the entry is
left unchanged. If the existing next hop is the same as the new one,
the metric is updated regardless of the metric value. If no
corresponding route is found, a new route entry is created.
Murakami & Maruyama Informational [Page 5]RFC 2174 MAPOS June 1997
3.4.2 Route Expiration
Assume a route to R is advertised by a neighboring switch S. If no
update message has been received from switch S for the period
FULL_UPDATE_TIME * 3 (30 seconds by default) or the route is
advertised with metric 16 by switch S, the route to R is marked as
unreachable by setting its metric to 16. In other words, the route to
R is kept advertised even if the route is not refreshed up-to 30
seconds.
To process this, each routing table entry has an EXPIRATION_TIMER (30
seconds by default, that is, FULL_UPDATE_TIME *3). If another switch
advertises a route to R, it replaces the unreachable route. Even if a
route is marked unreachable, the entry is kept in the routing table
for the period of FULL_UPDATE_TIME * 3. This enables the switch to
notify its neighbors of the unreachable route by sending update
messages with metric 16. To process this, each routing table entry
has a garbage collection timer GC_TIMER (30 seconds by default). The
entry is deleted on expiration of the timer. Figure 3 shows this
transition.
The Last Update Expiration Garbage Collection
| | |
Routing V T T T V T T T V
Table +-------+-------+-------+-------+-------+-------X
Entry metric < 16 | metric = 16 |
----------------------->|---------------------->|
EXPIRATION_TIMER GC_TIMER
Stop Advertising
|
Advertised V
Metric -- metric <16 ------+-- metric = 16 -------X
T: FULL_UPDATE_TIME
Figure 3. Route Expiration
3.4.3 Slow Convergence Prevention
To prevent slow convergence of routing information, two techniques,
split horizon with poisoned reverse, and triggered update are
employed.
Murakami & Maruyama Informational [Page 6]RFC 2174 MAPOS June 1997
Sn <------------- S3 <- S2 <- S1
(i) Before Outage
->
Sn <-- X -- S3 <- S2 <- S1
(ii) After Outage
Figure 4 An Example of Slow Convergence
Figure 4 shows an example of slow convergence[6]. In (i), three
switches, S1, S2, and S3, are assumed to have a route to Sn. In (ii),
the connection to Sn has disappeared because of an outage, but S2
continue to advertise the route since there is no means for S2 to
detect the outage immediately and it has the route to Sn in its
routing table. Thus, S3 misunderstand that S2 has the best route to
Sn and S2 is the next hop. This results in a transitive loop between
S2 and S3. S2 and S3 increments the metric of the route to Sn every
time they advertise the route and the loop continues until the metric
reaches 16. To suppress the slow convergence problem, split horizon
with poisoned reverse is used.
In split horizon with poisoned reverse, a route is advertised as
unreachable to the next hop. The metric is the received metric value
plus 16. For example, in Figure 4, S2 advertises the route to Sn with
the metric unreachable only to S3. Thus, S3 never considers that S2
is the next hop to Sn. This ensures fast convergence on disappearance
of a route.
Another technique, triggered update, forces a switch to send an
immediate update instead of waiting for the next periodic update when
a switch detects a local port failure, or when it receives a message
that a route has become unreachable, or that its metric has
increased. This makes the convergence faster.
4. Broadcast/multicast Routing in SSP
This section explains VRPB algorithm and the outline of
broadcast/multicast routing protocol.
Murakami & Maruyama Informational [Page 7]RFC 2174 MAPOS June 1997
4.1 Virtual Reverse Path Broadcast/Multicast Algorithm
SSP provides broadcast/multicast routing based on a spanning tree
algorithm. As described in Section 2, the routing is based on the
VRPB(Virtual Reverse Path Broadcast) algorithm. In VRPB, each switch
assumes that all broadcast and multicast frames are generated by a
specific switch, VSS(Virtual Source Switch). Thus, unlike DVMRP, a
MAPOS network has only one spanning tree at any given time.
The frames are forwarded along the reverse path by computing the
shortest path from the VSS to all possible recipients. VSS is the
switch which has the lowest switch number in the network. Because
the routing table contains all the unicast destination addresses
including the switch numbers, each switch can identify the VSS
independently by searching for the smallest switch number in its
unicast routing table.
In Figure 2, switch S1 is the VSS. Each switch determines its place
in the spanning tree, relative to the VSS, and which of its ports are
on the shortest path tree. Thus, the spanning tree is as shown in
Figure 5. Except for the VSS, each switch has one upstream port and
zero or more downstream ports. VSS have no upstream port, since it is
the root of the spanning tree. In Figure 2. switch S2's upstream
port is port 0x09 and it has no downstream port.
S1 (VSS)
/ \
/ \
/ \
S2 S3
Figure 5 VRPB Spanning Tree
When a switch receives a broadcast/multicast frame, it forwards the
frame to all of the upstream switch, the downstream switches, and the
directly connected nodes. However, it does not forward to the switch
which sent the frame to it. For that purpose, a bit mapped
broadcast/multicast routing table may be employed. The
broadcast/multicast routing process marks all the bits corresponding
to the ports to which frames should be forwarded. The forwarding
process refers to it and broadcasts a frame to all the ports with its
corresponding bit marked.
4.2 Forwarding Broadcast/multicast Frames
When a switch forwards a broadcast/multicast frame, (1) it first
decides the VSS by referring to its unicast routing table. Then, (2)
it refers to its broadcast/multicast routing table corresponding to
Murakami & Maruyama Informational [Page 8]RFC 2174 MAPOS June 1997
the VSS. A cache may be used to reduce the search overhead. (3) Based
on the routing table, the switch forwards the frame.
Figure 6 shows an example of S2's broadcast/multicast routing table
for the VSS S1. It is a bit map table and each bit corresponds to a
port. The value 1 indicates that frames should be forwarded to a node
or a switch through the port. If no bit is marked, the frame is
silently discarded. In the example of Figure 6, port 0x09 is the
upstream port to its VSS, that is, S1. Other ports, ports 0x05 and
0x03 are path to N2 and N1 nodes, respectively.
0F 0D 0B 09 07 05 03 01 --- port number
+---+---+---+---+---+---+---+---+
| 0 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | --- 1: forward
+---+---+---+---+---+---+---+---+ 0: inhibit
Figure 6 Broadcast/Multicast Routing Table of S2
4.3 Forwarding Path Examples
Assume that a broadcast frame is generated by N2 in Figure 2. The
frame is received by S2.
Then, S2 passes it to all the connected nodes except for the source
N2. That is, only to N1. At the same time, it also forwards the frame
to all its upstream and downstream switches. Since S2 has no
downstream switch, S2 forwards the frame to S1 though its upstream
port 0x09.
S1 is the VSS and it passes the frame to all the local nodes, that
is, only to N3. Since it has no upstream switch and S2 is the switch
which sent the frame to S1, the frame is eventually forwarded only to
a downstream switch S3.
S3 passes the frame to its local node, N4. Since S3 has only an
upstream and the frame was received through that port, S3 does not
forward the frame to any switch.
The resulting path is shown in Figure 7. Although this is not the
optimal path, VRPB ,at least, ensures that broadcast/multicast frames
are delivered all the nodes without a loop. Figures 8 and 9 show the
forwarding path for frames generated by a node under S3 and S4,
respectively.
Murakami & Maruyama Informational [Page 9]RFC 2174 MAPOS June 1997
+-> N3
|
N2 -> S2 +-> S1 +-> S3 -> N4
|
+-> N1
Figure 7 Forwarding Path from N2
+-> N1
|
N3 -> S1 +-> S2 +-> N2
|
+-> S3 --> N4
Figure 8 Forwarding Path from N3
+-> N3
|
N4 -> S3 +-> S1 +-> S2 +-> N1
|
+-> N2
Figure 9 Forwarding Path from N4
4.4 Suppressing Routing Loop
To suppress transitive routing loop, forward delay is employed. A
switch suspends broadcast/multicast forwarding for a period after a
new VSS is found in the routing table. This prevents transitive
routing loop by waiting for all the switches to have the same routing
information and become synchronized. In addition to controlling
sending of frames by forward delay, another mechanism is employed to
prevent transitive routing loop by controlling reception of frames.
That is, broadcast/multicast frames received through ports other than
the upstream and downstream ports are discarded.
4.5 Upstream Switch Discovery
The upstream port is determined by the shortest reverse path to the
VSS. It is identified by referring to the next hop port of the route
to VSS in the local unicast routing table. When a new next hop to the
VSS is discovered, the bit corresponding to the old next hop port is
cleared, and the bit corresponding to the new one is marked as the
upstream port in the broadcast/multicast routing table.
Murakami & Maruyama Informational [Page 10]RFC 2174 MAPOS June 1997
4.6 Downstream Switch Discovery
To determine the downstream ports, split horizon with poisoned
reverse is employed. When a switch receives a route with a metric
poisoned by split horizon processing through a port as described in
Section 3.4.3, the port is considered to be a downstream port. In
Figure 2, S1 is the VSS and the route information is sent back from
S2 to S1 with metric unreachable based on the split horizon with
poisoned reverse. Thus, S1 knows that S2 is one of its downstreams.
4.7 Downstream Port Expiration
When a poison reversed packet is newly received from a port, the
local switch knows that a new downstream switch has appeared. Then,
it marks the bit corresponding to the port and starts
FORWARD_DELAY_TIMER (30second by default, that is, FULL_UPDATE_TIME *
3) for the port. The forwarding of broadcast/multicast frames to the
port is prohibited until the timer expires. Every time the local
switch receives a poison reversed packet through a port, it
initializes PORT_EXPIRATION_TIMER(30 seconds by default, that is,
FULL_UPDATE_TIME *3) corresponding to the port. A continuous loss of
poison reversed packets or a failure of downstream port results in
expiration of PORT_EXPIRATION_TIMER, and the corresponding bit is
cleared.
First Update Last Update
| |
V T T T T T T T V
+---+---+---+---+---+---+---+---+---+---+---+---+---
A bit in
the routing 0 0 0 1 1 1 1 1 1 1 0 0 0
table ^ ^
<--------->| <--------->|
^ route up ^ route down
| |
FORWARD_DELAY PORT_EXPIRATION
T: FULL_UPDATE_TIME
Figure 10. Port Expiration
When a downstream switch discovers another best path to the VSS or a
new VSS, it stops split horizon with poison reverse and sends
ordinary update messages. Whenever the local switch receives an
ordinary update message from its downstream switch, it SHOULD
immediately clear the corresponding bit in the routing table and stop
forwarding of broadcast/multicast frames.
Murakami & Maruyama Informational [Page 11]RFC 2174 MAPOS June 1997
4.8 Node Discovery
When a NSP[9] packet, requesting a node address from a port, is
received, the local switch considers that a new node is connected,
and marks the corresponding bit in the broadcast/multicast routing
table. When the local switch detects that the port went down as
described in [9], it clear the corresponding bit.
4.9 Invalidating The Broadcast/multicast Routing Table
When a new VSS is discovered or when the VSS becomes unreachable, the
entire broadcast/multicast routing table is invalidated. That is, a
change of upstream port affects the entire broadcast/multicast
routing. However, a change of a downstream port does not affect
forwarding to other downstream ports, its upstream port, and nodes.
5. Detailed Protocol Operation
This section explains SSP packet format and protocol processing in
detail.
5.1 Packet Format
This subsection describes the packet encapsulation in HDLC frame and
the packet format.
5.1.1 Packet Format and Its Encapsulation
SSP packet format is designed based on RIP[6] and its successor, RIP2
[7]. Figure 11 shows the packet format. A SSP packet is encapsulated
in the information field of a MAPOS HDLC frame. The HDLC protocol
field of SSP is 0xFE05 in hex as defined by the "MAPOS Version 1
Assigned Numbers" [10]. The packet is sent encapsulated in a unicast
packet with the destination address 0000 0001, which indicates the
control processor of an adjacent switch.
Murakami & Maruyama Informational [Page 12]
RFC 2174 MAPOS June 1997
(MSB) (LSB)
7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ -----
| Command | Version | unused |SSP header
+---------------+---------------+-------------------------------+ -----
| Address Family Identifier | All 0 |
+-------------------------------+-------------------------------+
| HDLC Address | an SSP
+---------------------------------------------------------------+ route
| Subnet Mask | entry
+---------------------------------------------------------------+
| All 0 |
+---------------------------------------------------------------+
| Metric |
+---------------+---------------+-------------------------------+ ----
| Address Family Identifier | All 0 |
Figure 11 SSP packet format
The maximum packet size is 512 octet. The first four octets is the
SSP header. The remainder of the message is composed of 1 - 25 route
entries. Each entry is 20 octets long.
5.1.2 SSP Header
SSP header consists of a command field and a version field. The
command field is one octet long and holds one of the following
values;
1 - request A request to send all or part of SSP routing table.
2 - response A message containing all, or a part of the sender's
SSP routing table. This message may be sent in
response to a request, or it may be an update
message generated by the sender.
The Version field indicates the version of SSP being used. The
current version number is 1.
5.1.3 SSP Route Entries
Each entry has an address family identifier. It indicates an
attribute of the entry. SSP routing protocol uses 2 as its identifier
by default. The identifier 0 indicates unspecified. This value is
used when a switch requests other switches to send the entire SSP
routing table. A recipient of the message SHOULD ignore all entries
with unknown value.
Murakami & Maruyama Informational [Page 13]RFC 2174 MAPOS June 1997
The HDLC address is a destination address. It may be a switch address
or a node address. The subsequent subnet mask is applied to the HDLC
address to yield the switch number portion. The field is 4 octet long
and the address is placed in the least significant position.
Metric indicates the distance to the destination node. That is, how
many switches a message must go through en route to the destination
node. The metric field must contain a value between 1 and 31. The
metric of 16 indicates that the destination is not reachable and is
ignored by recipients. The values between 17 and 31 are utilized for
poisoned reverse with split horizon and also means unreachable. The
metric 0 indicates the local switch itself.
5.2 Routing Table
Every switch has an SSP routing table. The table is a collection of
route entries - one for every destination. An entry consists of the
following information;
(1) destination : A unicast destination address.
(2) subnet mask : A mask to extract the switch address by applying
bitwise AND with the destination address
(3) next hop port : The local port number connected to the adjacent
switch along the path to the destination.
(4) metric : Distance to the destination node. The metric of an
adjacent switch is 1 and that of local switch is 0.
(5) timers for unicast routing : Timers associated with unicast
routing such as EXPIRATION_TIMER and GC_TIMER.
(6) flags : Various flags associated with the route such as route
change flag to indicate that the route has changed recently or it
has timed out.
(7) bit map routing table for broadcast/multicast : Each bit
corresponding to the port to an upstream or a downstream switch of
the spanning tree is marked in addition to the ports to end nodes.
Broadcast/multicast frames are forwarded only through those ports
with their corresponding bit set. Since only one spanning tree
exists at a time in a network, each route entry does not necessarily
have to have this field.
Murakami & Maruyama Informational [Page 14]RFC 2174 MAPOS June 1997
(8) timers for broadcast/multicast routing : Timers associated with
broadcast/multicast routing such as FORWARD_DELAY_TIMER and
PORT_EXPIRATION_TIMER. These timers are prepared for each bit of
broadcast/multicast routing table.
5.3 Sending Routing Messages
5.3.1 Packet Construction
Because of the split horizon with poisoned reverse, a routing message
differs depending on the adjacent switch to which the message is
being sent. The upstream switch of a route, that is next hop,
receives a message which contains the corresponding route with a
metric between 17 and 31. Switches that are not the upstream switch
of any route receive the same message. Here, we assume that a packet
for a routing message is constructed for an adjacent switch which is
connected through the local port N.
First, set the version field to 1, the current SSP version. Then, set
the command to "response". Set other fields which are supposed to be
zero to zero. Next, start filling in entries.
To fill in the entries, perform the following for each route. The
destination HDLC address, netmask, and its metric are put into the
entry in the packet. Routes must be included in the packet even if
their metrics are unreachable(16). If the next hop port is N, 16 is
added to the metric for split horizon with poisoned reverse.
Recall that the maximum packet size is 512 bytes. When there is no
more space in a packet, send the current message and start a new one.
If a triggered update is being generated, only entries whose route
change flags are set need be included.
5.3.2 Sending update
Sending update may be triggered in any of the following ways;
(1) Initial Update
When a switch first comes up, it SHOULD send to all adjacent
switches a request asking for their entire routing tables. The
destination address is 00000001. When a port comes on-line, the
request packet is sent to the port. The packet, requesting the
entire routing table, MUST have at least an entry with the address
family identifier 0 meaning unspecified.
When a switch receives a request packet, it first checks the version
number of the SSP header. If it is not 1, the packet is silently
Murakami & Maruyama Informational [Page 15]RFC 2174 MAPOS June 1997
discarded. Otherwise, the address family identifier is examined. If
the value is 0, the entire SSP routing table is returned in one or
more response packets destined to 00000001. Otherwise, the request
is silently discarded. Although the original RIP specification
defines the partial routing table request, SSP routing protocol
omits it for the sake of simplicity.
(2) Periodic Update
Every switch participating in the routing process sends an update
message (response message) to all its neighbor switches once every
FULL_UPDATE_TIME (10 seconds). For the periodic update, a response
packet(s) is used. The destination address is always 00000001. An
update message contains the entire SSP routing table. The maximum
packet size is 512byte. Thus, an update message may require several
packets to be packed.
(3) Triggered Update
When a route in the unicast routing table is changed or a local port
goes down, the switch advertises a triggered update packet without
waiting for the full update time. The difference between triggered
update and the other update is that triggered updates do not have to
include the entire routing table. Only changed entries should be
included. Triggered update may be suppressed if a regular periodic
update is due.
Note that when a route is advertised as unreachable (metric 16) by
an adjacent switch, update process is triggered as well as
expiration of the route in the local switch.
(4) On Termination
When a switch goes down, it is desirable to advertise all the routes
with metric 16, that is, unreachable.
5.4 Receiving Routing Messages
When a switch receives an update, it first checks the version number.
If it is not 1, the update packet is silently discarded. Otherwise,
it processes the entries in it one by one.
Murakami & Maruyama Informational [Page 16]RFC 2174 MAPOS June 1997
For each entry, the address family identifier is checked. If it is
not 2, the entry is ignored. Otherwise, the metric is checked. The
value should be between 0 and 31. An entry with illegal metric is
ignored. Next, the HDLC address and the subnet mask is checked. An
entry with an invalid address such as broadcast is ignored. If the
entry passed all these validation checks, it is processed according
to the following steps;
Step 1 - Process Poisoned Reverse
If the metric value is between 0 and 16, it is an unicast
information. Go ahead to Step 2.
If the metric value is between 17 and 31, it indicates poisoned
reverse, that the local switch has been chosen as the next hop for
the route. However, if the corresponding entry is not included in the
current routing table or the message is from a port connected to its
upstream switch, the message is illegal -- ignore it and return to
Step 1 to process the next entry. Otherwise,
(1) Initialize the PORT_EXPIRATION_TIMER corresponding to the
downstream port.
(2) Operate the FORWARD_DELAY_TIMER as follows;
(2-1) If the broadcast/multicast forwarding was already
enabled, go to (3).
(2-2) If the FORWARD_DELAY_TIMER corresponding to the
downstream port was already started, increment the
timer. If the timer expires, mark the bit in the
broadcast/multicast routing table corresponding to the
port and stop the timer.
(2-2) Otherwise, start the FORWARD_DELAY_TIMER.
(3) Return to Step 1 to process the next entry.
Step 2 - Process Unicast Routing Information
First, add the cost associated with the link, usually 1, to the
metric. If the result is greater than 16, 16 is used. Then, look up
the unicast routing table for the corresponding entry. There are two
cases.
Case 1 no corresponding entry is found
If the new metric is 16, return to step 1 to process the next
entry. Otherwise,
(1) Create a new route entry in the routing table
(2) Initialize EXPIRATION_TIMER and GC_TIMER
Murakami & Maruyama Informational [Page 17]RFC 2174 MAPOS June 1997
(3) The port corresponding to the new route is the next_hop port
for the route. Thus, mark the bit in the broadcast/multicast
routing table corresponding to the new next_hop port and
start FORWARD_DELAY_TIMER. If this new route is for the
switch with the minimum switch number, select it as the VSS
and use its broadcast/multicast routing table. (See NOTE 1.)
(4) Set the route change flag and invoke triggered update process
(5) Return to step 1 to process the next entry.
[NOTE 1]
There are two implementations;
(1) Prepare a spanning tree for each route and use
only one corresponding to the current VSS. In this
case, each unicast route entry has a broadcast/unicast
routing table.
(2) Prepare only one spanning tree corresponding to the
current VSS. In this case, a switch has only one
broadcast/multicast routing table.
In this document, the former is assumed.
Case 2. A corresponding entry is found
In this case, the update message is processed differently
according to the new metric value.
(a) new_metric < 16 & new_metric > current_metric
(1)If and only if the update is from the same port(next_hop
port) as the existing one,
(1-1) Update the entry
(1-2) Initialize EXPIRATION_TIMER and GC_TIMER
(2) If the corresponding bit to the port, which the update
message is received, is marked in the broadcast/multicast
routing table, clear the bit.
(3) Return to Step 1 and process the next entry.
(b) new_metric < 16 & new_metric < current_metric
(1) Update the entry and clear the bit in the
broadcast/multicast routing table corresponding to the old
next_hop port.
(2) Initialize EXPIRATION_TIMER, GC_TIMER, and
PORT_EXPIRATION_TIMER for the new next_hop port.
(3) Mark a bit in the broadcast/multicast routing table
corresponding to the new next_hop port and start
FORWARD_DELAY_TIMER.
Murakami & Maruyama Informational [Page 18]RFC 2174 MAPOS June 1997
(4) Set the route change flag and invoke triggered update with
poisoned reverse for the new next_hop.
(5) Return to Step 1 to process the next entry.
(c) new_metric < 16 & new_metric = current_metric
If a new route with the same metric value as the existing
routing table entry is received, use the old one as follows;
(1) If the new next hop is equal to the current one,
initialize EXPIRATION_TIMER and GC_TIMER. Otherwise,
ignore this update.
(2) If the bit corresponding to the port, from which the
update message was received, is marked in the
broadcast/multicast routing table, clear the bit.
(3) Return to Step 1 to process the next entry.
(d) the new metric = 16 & the new next hop = the current one
If the current metric is not equal to 16, this is a new
unreachable information. Then,
(1) Update the entry and clear the bit in the
broadcast/multicast routing table corresponding to the old
next_hop port.
(2) If this route is for the current VSS, select a new VSS in
the valid routing table entries. Valid means that the
destination is reachable.
(3) Set the route change flag and invoke triggered update
process to notify the unreachable route.
Otherwise,
do nothing and return to Step 1 to process the next entry.
(e) the new metric = 16 & the new next hop /= the current one
(1) If the bit corresponding to the port, from which the
update message was received, is marked in the
broadcast/multicast routing table, clear the bit.
(2) Return to Step 1 to process the next entry.
Murakami & Maruyama Informational [Page 19]RFC 2174 MAPOS June 1997
5.5 Timers
The timer routine increments the following timers and executes its
associated process on their expiration.
(1) EXPIRATION_TIMER and GC_TIMER
The EXPIRATION_TIMERs and GC_TIMERs of each entry in the unicast
routing table are incremented every FULL_UPDATE_TIME (10 seconds by
default). When a EXPIRATION_TIMER expires, the metric is changed to
unreachable(16), update process is triggered, and GC_TIMER is
started. When a GC_TIMER expires, the entry is deleted from the
local routing table. EXPIRATION_TIMER and GC_TIMER are cleared every
time a switch receives a routing update.
(2) FORWARD_DELAY_TIMER
FORWARD_DELAY_TIMER is completely handled in the receive process and
has no relation to the timer routine.
(3) PORT_EXPIRATION_TIMER
PORT_EXPIRATION_TIMERs associated with each bit in the
broadcast/multicast routing table are incremented every
FULL_UPDATE_TIME (10 seconds by default). When the timer expires,
the corresponding downstream switch is considered to be down and the
corresponding bit in the broadcast/multicast routing table is
cleared. This timer is cleared by the receive process every time a
poisoned reverse packet is received from the corresponding switch.
6. Further considerations on implementation
6.1 Port State
A switch assumes that every port is connected to a switch initially.
Thus, it sends update packets to every port. When a node is connected
to a port, the switch recognizes it by receiving an NSP request
packet, and stops sending SSP packets to the port. Whenever a switch
detects a connection failure such as loss of signal and out-of-
synchronization, it should clear the internal state table
corresponding of the port.
6.2 Half way connection problem
A port consists of two channels, transmit and receive. Although it is
easy for a node or a switch to detect a receive channel failure,
transmit channel failure may not be detected, causing half way
connection. This results in a black hole.
Murakami & Maruyama Informational [Page 20]RFC 2174 MAPOS June 1997
Thus, whenever a switch receives a SSP update packet from a port, it
SHOULD check the status of the corresponding transmit channel.
SONET/SDH has a feedback mechanism for that purpose. The status of
the local transmit channel received at the remote end can be sent
back utilizing the overhead part, FEBE(Far End Block Error) and
FERF(Far End Receive Failure), of the corresponding receive channel.
If the signals indicates that the transmit channel has a problem, the
SSP packet received from the remote end should be silently discarded.
However, some SONET/SDH services do not provide path overhead
transparency.
Although, SONET/SDH APS(Automatic Protection Switching) can be
utilized to switch service from a failed line to a spare line, the
function is out of scope of this protocol.
7. Security Considerations
Security issues are not discussed in this memo.
References
[1] Murakami, K. and M. Maruyama, "MAPOS - Multiple Access Protocol
over SONET/SDH Version 1," RFC2171, June 1997.
[2] CCITT Recommendation G.707: Synchronous Digital Hierarchy Bit
Rates, 1990.
[3] CCITT Recommendation G.708: Network Node Interface for
Synchronous Digital Hierarchy, 1990.
[4] CCITT Recommendation G.709: Synchronous Multiplexing Structure,
1990.
[5] American National Standard for Telecommunications - Digital
Hierarchy - Optical Interface Rates and Formats Specification,
ANSI T1.105-1991.
[6] Hedrick, C., "Routing Information Protocol", STD 34, RFC 1058,
Rutgers University, June 1988.
[7] Malkin, G., "RIP Version 2 - Carrying Additional Information ",
RFC1723, Xylogics, Inc., November 1994.
[8] Pusateri, T., "Distance Vector Multicast Routing Protocol",
September 1996, Work in Progress.
[9] Murakami, K. and M. Maruyama, "A MAPOS version 1 Extension -
Node Switch Protocol," RFC2173, June 1997.
Murakami & Maruyama Informational [Page 21]RFC 2174 MAPOS June 1997
[10] Maruyama, M. and K. Murakami, "MAPOS Version 1 Assigned
Numbers," RFC2172, June 1997.
Acknowledgements
The authors would like to acknowledge the contributions and
thoughtful suggestions of John P. Mullaney, Clark Bremer, Masayuki
Kobayashi, Paul Francis, Toshiaki Yoshida, Takahiro Sajima, and
Satoru Yagi.
Authors' Address
Ken Murakami
NTT Software Laboratories
3-9-11, Midori-cho
Musashino-shi
Tokyo 180, Japan
E-mail: murakami@ntt-20.ecl.net
Mitsuru Maruyama
NTT Software Laboratories
3-9-11, Midori-cho
Musashino-shi
Tokyo 180, Japan
E-mail: mitsuru@ntt-20.ecl.net
Murakami & Maruyama Informational [Page 22]
