INTERNET-DRAFT Linda Dunbar
Intended status: Proposed Standard Donald Eastlake
Huawei
Radia Perlman
EMC
Igor Gashinsky
Yahoo
Yizhou Li
Huawei
Expires: May 9, 2014 November 10, 2014
TRILL: Edge Directory Assist Mechanisms
<draft-ietf-trill-directory-assist-mechanisms-01.txt>
Abstract
This document describes mechanisms for providing directory service to
TRILL (Transparent Interconnection of Lots of Links) edge switches.
The directory information provided can be used in reducing multi-
destination traffic, particularly ARP/ND and unknown unicast
flooding.
Status of This Memo
This Internet-Draft is submitted to IETF in full conformance with the
provisions of BCP 78 and BCP 79.
Distribution of this document is unlimited. Comments should be sent
to the TRILL working group mailing list.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/1id-abstracts.html. The list of Internet-Draft
Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.
L. Dunbar, et al [Page 1]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Table of Contents
1. Introduction............................................3
1.1 Uses of Directory Information..........................3
1.2 Terminology............................................4
2. Push Model Directory Assistance Mechanisms..............6
2.1 Requesting Push Service................................6
2.2 Push Directory Servers.................................6
2.3 Push Directory Server State Machine....................7
2.3.1 Push Directory States................................7
2.3.2 Push Directory Events and Conditions.................8
2.3.3 State Transition Diagram and Table...................9
2.4 Additional Push Details...............................10
2.5 Primary to Secondary Server Push Service..............11
3. Pull Model Directory Assistance Mechanisms.............12
3.1 Pull Directory Message Common Format..................13
3.2 Pull Directory Query and Response Messages............14
3.2.1 Pull Directory Query Message Format.................14
3.2.2 Pull Directory Response Format......................17
3.3 Cache Consistency.....................................19
3.3.1 Update Message Format...............................21
3.3.2 Acknowledge Message Format..........................22
3.4 Pull Directory Hosted on an End Station...............22
3.5 Pull Directory Message Errors.........................24
3.6 Additional Pull Details...............................25
4. Directory Use Strategies and Push-Pull Hybrids.........27
4.1 Strategy Configuration................................27
5. Security Considerations................................30
6. IANA Considerations....................................31
6.1 ESADI-Parameter Data Extensions.......................31
6.2 RBridge Channel Protocol Number.......................32
6.3 The Pull Directory (PUL) and No Data (NOD) Bits.......32
Acknowledgments...........................................34
Normative References......................................35
Informational References..................................36
Authors' Addresses........................................37
L. Dunbar, et al [Page 2]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
1. Introduction
[RFC7067] gives a problem statement and high level design for using
directory servers to assist TRILL [RFC6325] edge nodes in reducing
multi-destination ARP/ND, reducing unknown unicast flooding traffic,
and improving security against address spoofing within a TRILL
campus. Because multi-destination traffic becomes an increasing
burden as a network scales up in number of nodes, reducing ARP/ND and
unknown unicast flooding improves TRILL network scalability. This
document describes specific mechanisms for directory servers to
assist TRILL edge nodes. These mechanisms are optional to implement.
The information held by the Directory(s) is address mapping and
reachability information. Most commonly, what MAC address [RFC7042]
corresponds to an IP address within a Data Label (VLAN or FGL (Fine
Grained Label [RFC7172])) and the egress TRILL switch (RBridge), and
optionally what specific TRILL switch port, from which that MAC
address is reachable. But it could be what IP address corresponds to
a MAC address or possibly other address mappings or reachability.
In the data center environment, it is common for orchestration
software to know and control where all the IP addresses, MAC
addresses, and VLANs/tenants are in a data center. Thus such
orchestration software can be appropriate for providing the directory
function or for supplying the Directory(s) with directory
information.
Directory services can be offered in a Push or Pull Mode [RFC7067].
Push Mode, in which a directory server pushes information to TRILL
switches indicating interest, is specified in Section 2. Pull Mode,
in which a TRILL switch queries a server for the information it
wants, is specified in Section 3. More detail on modes of operation,
including hybrid Push/Pull, are provided in Section 4.
The mechanism used to initially populate directory data in primary
servers is beyond the scope of this document. A primary server can
use the Push Directory service to provide directory data to secondary
servers as described in Section 2.5.
1.1 Uses of Directory Information
A TRILL switch can consult Directory information whenever it wants,
by (1) searching through information that has been retained after
being pushed to it or pulled by it or (2) by requesting information
from a Pull Directory. However, the following are expected to be the
most common circumstances leading to directory information use. All
of these are cases of ingressing (or originating) a native frame.
L. Dunbar, et al [Page 3]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
1. ARP requests and replies [RFC826] are normally broadcast. But a
directory assisted edge TRILL switches could intercept ARP
messages and reply if the TRILL switch has the relevant
information.
2. IPv6 ND (Neighbor Discovery [RFC4861]) requests and replies are
normally multicast. Except in the case of Secure ND [RFC3971]
where possession of the right keying material might be required,
directory assisted edge TRILL switches could intercept ND messages
and reply if the TRILL switch has the relevant information.
3. Unknown destination MAC addresses. An edge TRILL switch ingressing
a native frame necessarily has to determine if it knows the egress
RBridge from which the destination MAC address of the frame (in
the frame's VLAN or Fine Grained Label) is reachable. It might
learn that information from the directory or could query the
directory if it does not know. Furthermore, if the edge TRILL
switch has complete directory information, it can detect forged
source MAC address on the native frame and discard the frame in
that case.
4. RARP [RFC903] is similar to ARP as above.
1.2 Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119 [RFC2119].
The terminology and acronyms of [RFC6325] are used herein along with
the following:
COP: Complete Push flag bit. See Sections 2 and 6.1 below.
CSNP Time: Complete Sequence Number PDU Time. See ESDADI [RFC7357]
and Section 6.1 below.
Data Label: VLAN or FGL.
FGL: Fine Grained Label [RFC7172].
Host: Application running on a physical server or a virtual machine.
A host must have a MAC address and usually has at least one IP
address.
IP: Internet Protocol. In this document, IP includes both IPv4 and
IPv6.
L. Dunbar, et al [Page 4]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
PSH: Push Directory flag bit. See Sections 2 and 6.1 below.
PUL: Pull Directory flag bit. See Sections 3 and 6.3 below.
primary server: A Directory server that obtains the information it is
serving up by a reliable mechanism outside the scope of this
document designed to assure the freshness of that information.
(See secondary server.)
RBridge: An alternative name for a TRILL switch.
secondary server: A Directory server that obtains the information it
is serving up from one or more primary servers.
tenant: Sometimes used as a synonym for FGL.
TRILL switch: A device that implements the TRILL protocol.
L. Dunbar, et al [Page 5]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
2. Push Model Directory Assistance Mechanisms
In the Push Model [RFC7067], one or more Push Directory servers
reside at TRILL switches and push down the address mapping
information for the various addresses associated with end station
interfaces and the TRILL switches from which those interfaces are
reachable [IA]. This service is scoped by Data Label (VLAN or FGL
[RFC7172]). A Push Directory also advertises whether or not it
believes it has pushed complete mapping information for a Data Label.
It might be pushing only a subset of the mapping and/or reachability
information for a Data Label. The Push Model uses the ESADI [RFC7357]
protocol as its distribution mechanism.
With the Push Model, if complete address mapping information for a
Data Label being pushed is available, a TRILL switch (RBridge) which
has that complete pushed information and is ingressing a native frame
can simply drop the frame if the destination unicast MAC address
can't be found in the mapping information available, instead of
flooding the frame (ingressing it as an unknown MAC destination TRILL
Data frame). But this will result in lost traffic if ingress TRILL
switch's directory information is incomplete.
2.1 Requesting Push Service
In the Push Model, it is necessary to have a way for a TRILL switch
to request information from the directory server(s). TRILL switches
simply use the ESADI [RFC7357] protocol mechanism to announce, in
their core IS-IS LSPs, the Data Labels for which they are
participating in ESADI by using the Interested VLANs and/or
Interested Labels sub-TLVs [RFC7176]. This will cause them to be
pushed the Directory information for all such Data Labels that are
being served by one or more Push Directory servers.
2.2 Push Directory Servers
Push Directory servers advertise their availability to push the
mapping information for a particular Data Label to each other and to
ESADI participants for that Data Label through ESADI by turning on
the a flag bit in their ESADI Parameter APPsub-TLV for that ESADI
instance (see [RFC7357] and Section 6.1). Each Push Directory server
MUST participate in ESADI for the Data Labels for which it will push
mappings and set the PSH (Push Directory) bit in its ESADI-Parameters
APPsub-TLV for that Data Label.
For robustness, it is useful to have more than one copy of the data
being pushed. Each Push Directory server is configured with a number
L. Dunbar, et al [Page 6]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
N in the range 1 to 8, which defaults to 2, for each Data Label for
which it can push directory information. If the Push Directories for
a Data Label are configured the same in this regard and enough such
servers are available, N copies of the directory that will be pushed.
Each Push Directory server also has an 8-bit priority to be Active
(see Section 6.1 of this document). This priority is treated as an
unsigned integer where larger magnitude means higher priority and is
in its ESADI Parameter APPsub-TLV. In cases of equal priority, the
6-byte IS-IS System IDs of the tied Push Directories are used as a
tie breaker and treated as an unsigned integer where larger magnitude
means higher priority.
For each Data Label it can serve, each Push Directory server orders,
by priority, the Push Directory servers that it can see in the ESADI
link state database for that Data Label that are data reachable
[RFC7180] and determines its own position in that order. If a Push
Directory server is configured to believe that N copies of the
mappings for a Data Label should be pushed and finds that it is
number K in the priority ordering (where number 1 is highest priority
and number K is lowest), then if K is less than or equal to N the
Push Directory server is Active. If K is greater than N it is
Passive. Active and Passive behavior are specified below.
For a Push Directory to reside on an end station, one or more TRILL
switches locally connected to that end station must proxy for the
Push Directory server and advertise themselves as Push Directory
servers. It appears to the rest of the TRILL campus that these TRILL
switches (that are proxying for the end station) are the Push
Directory server(s). The protocol between such a Push Directory end
station and the one or more proxying TRILL switches acting as Push
Directory servers is beyond the scope of this document.
2.3 Push Directory Server State Machine
The subsections below describe the states, events, and corresponding
actions for Push Directory servers.
2.3.1 Push Directory States
A Push Directory Server is in one of six states, as listed below, for
each Data Label it can serve. In addition, it has an internal State-
Transition-Time variable for each Data Label it can serve which is
set at each state transition and which enables it to determine how
long it has been in its current state for that Data Label.
L. Dunbar, et al [Page 7]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Down: A completely shut down virtual state defined for convenience in
specifying state diagrams. A Push Directory Server in this state
does not advertise any Push Directory data. It may be
participating in ESDADI [RFC7357] with the PSH bit zero in its
ESADI-Parameters or might be not participating in ESADI at all.
All states other than the Down state are considered to be Up
states.
Passive: No Push Directory data is advertised. Any outstanding EASDI-
LSP fragments containing directory data are updated to remove that
data and if the result is an empty fragment (contains nothing
except possibly an Authentication TLV), the fragment is purged.
The Push Directory participates in ESDADI [RFC7357] and advertises
its ESADI fragment zero that includes an ESADI-Parameters APPsub-
TLV with the PSH bit set to one and COP (Complete Push) bit zero.
Active: If a Push Directory server is Active, it advertises its
directory data and any changes through ESADI [RFC7357] in its
ESADI-LSPs using the Interface Addresses [IA] APPsub-TLV and
updates that information as it changes. The PSH bit is set to one
in the ESADI-Parameters and the COP bit set to zero.
Completing: Same behavior as the Active state but responds
differently to events.
Complete: The same behavior as Active except that the COP bit in the
ESADI-Parameters APPsub-TLV is set to one and the server responds
differently to events.
Reducing: The same behavior as Complete but responds differently to
events. The PSH bit remains a one but the COP bit is cleared to
zero in the ESADI-Parameters APPsub-TLV. Directory updates
continue to be advertised.
2.3.2 Push Directory Events and Conditions
Three auxiliary conditions referenced later in this section are
defined as follows for convenience:
The Activate Condition: The Push Directory server determines that it
is priority K among the data reachable Push Directory servers
(where highest priority is 1), the server is configured that there
should be N copies pushed, and K is less than or equal to N. For
example, the Push Directory server is configured that 2 copies
should be pushed and finds that it is priority 1 or 2 among the
Push Directory servers it can see.
The Pacify Condition: The Push Directory server determines that it is
L. Dunbar, et al [Page 8]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
priority K among the data reachable data reachable Push Directory
servers (where highest priority is 1), the server is configured
that there should be N copies pushed, and K is greater than N. For
example, the Push Directory server is configured that 2 copies
should be pushed and finds that it is priority 3 or lower priority
(higher number) among the Push directory servers it can see.
The Time Condition: The Push Directory server has been in its current
state for an amount of time equal to or larger than its CSNP time
(see Section 6.1).)
The events and conditions listed below cause state transitions in
Push Directory servers.
1. Push Directory server was Down but is now up.
2. The Push Directory server or the TRILL switch on which it resides
is being shut down.
3. The Activate Condition is met and the server is not configured to
believe it has complete data.
4. The Pacify Condition is met.
5. The Activate Condition is met and the server is configured to
believe it has complete data.
6. The server is configured to believe it does not have complete
data.
7. The Time Condition is met.
2.3.3 State Transition Diagram and Table
The state transition table is as follows:
Event || Down |Passive |Active |Completing|Complete|Reducing|
------++-------+----------+--------+----------+--------+--------+
1 ||Passive|Passive |Active |Completing|Complete|Reducing|
2 || Down | Down |Passive |Passive |Reducing|Reducing|
3 || Down |Active |Active |Active |Reducing|Reducing|
4 || Down |Passive |Passive |Passive |Reducing|Reducing|
5 || Down |Completing|Complete|Completing|Complete|Complete|
6 || Down |Passive |Active |Active |Reducing|Reducing|
7 || Down |Passive |Active |Complete |Complete|Active |
The above state table is equivalent to the following transition
diagram:
L. Dunbar, et al [Page 9]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
+-----------+
| Down |<---------+
+-----------+ |
|1 ^ | 3,4,5,6,7 |
| | +------------+
V |2
+-----------+
| Passive |<-----------------------
+-----------+ ^ ^ ^
|5 |3 |1,4,6,7 | | |
| | +---------+ | |
| V |2,4 |
| +---------------------+ |
| | Active |<--+ |
| +---------------------+ | |
| |5 ^ |1,3,6,7 ^ | |
| | | | | | |
| | | +---------+ | |
| | | | |
V V |3,6 | |
+--------------+ | |
| Completing |-------------------+
+--------------+ 2,4 |
|7 |1,5 ^ |
| | | |
| +-----+ |
V |7
+-------------+ +----------------+
| Complete |--------->| Reducing |<--+
+-------------+ 2,3,4,6 +----------------+ |
|1,5,7 ^ ^ |5 |1,2,3,4,6 |
| | | | | |
+------+ +--------------+ +--------------+
Figure 1. Push Server State Diagram
2.4 Additional Push Details
Push Directory mappings can be distinguished for other data
distributed through ESADI because mappings are distributed only with
the Interface Addresses APPsub-TLV [IA] and are flagged as being Push
Directory data.
TRILL switches, whether or not they are a Push Directory server, MAY
continue to advertise any locally learned MAC attachment information
in ESDADI [RFC7357] using the Reachable MAC Addresses TLV [RFC6165].
However, if a Data Label is being served by complete Push Directory
servers, advertising such locally learned MAC attachment generally
L. Dunbar, et al [Page 10]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
SHOULD NOT be done as it would not add anything and would just waste
bandwidth and ESADI link state space. An exception might be when a
TRILL switch learns local MAC connectivity and that information
appears to be missing from the directory mapping.
Because a Push Directory server needs to advertise interest in one or
more Data Labels even if it does not want to receive end station
multidestination data in those Data Labels, the No Data (NOD) flag
bit is provided as specified in Section 6.3.
When a Push Directory server is no longer data reachable [RFC7180],
TRILL switches MUST ignore any Push Directory data from that server
because it is no longer being updated and may be stale.
The nature of dynamic distributed asynchronous systems is such that
it is impossible for a TRILL switch receiving Push Directory
information to be absolutely certain that it has complete
information. However, it can obtain a reasonable assurance of
complete information by requiring two conditions to be met:
1. The PSH and COP bits are on in the ESADI zero fragment from the
server for the relevant Data Label.
2. It has had continuous data connectivity to the server for the
larger of the client's and the server's CSNP times.
Condition 2 is necessary because a client TRILL switch might be just
coming up and receive an EASDI LSP meeting the requirement in
condition 1 above but have not yet received all of the ESADI LSP
fragment from the Push Directory server.
There may be conflicts between mapping information from different
Push Directory servers or conflicts between locally learned
information and information received from a Push Directory server. In
case of such conflicts, information with a higher confidence value
[RFC6325] is preferred over information with a lower confidence. In
case of equal confidence, Push Directory information is preferred to
locally learned information and if information from Push Directory
servers conflicts, the information from the higher priority Push
Directory server is preferred.
2.5 Primary to Secondary Server Push Service
A secondary Push or Pull Directory server is one that obtains its
data from a primary directory server. Other techniques MAY be used
but, by default, this data transfer occurs through the primary server
acting as a Push Directory server for the Data Labels involved while
the secondary directory server takes the pushed data it receives from
the highest priority Push Directory server and re-originates it. Such
a secondary server may be a Push Directory server or a Pull Directory
server or both for any particular Data Label.
L. Dunbar, et al [Page 11]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
3. Pull Model Directory Assistance Mechanisms
In the Pull Model [RFC7067], a TRILL switch (RBridge) pulls directory
information from an appropriate Directory Server when needed.
Pull Directory servers for a particular Data Label X are found by
looking in the core TRILL IS-IS link state database for data
reachable TRILL switches that advertise themselves by having the Pull
Directory flag (PUL) on in their Interested VLANs or Interested
Labels sub-TLV [RFC7176] for that Data Label. If multiple such TRILL
switches indicate that they are Pull Directory Servers for a
particular Data Label, pull requests can be sent to any one or more
of them but it is RECOMMENDED that pull requests be preferentially
sent to the server or servers that are lower cost from the requesting
TRILL switch.
Pull Directory requests are sent by enclosing them in an RBridge
Channel [RFC7178] message using the Pull Directory channel protocol
number (see Section 6.2). Responses are returned in an RBridge
Channel message using the same channel protocol number. See Section
3.2 for Query and Response message formats. For cache consistency or
notification purposes, Pull Directory servers can sent unsolicited
Update messages to client TRILL switches they believe may be holding
old data and those clients can acknowledge such updates, as described
in Section 3.3. All these messages have a common header as described
in Section 3.1. Errors returns can be sent for queries or updates as
described in Section 3.5.
The requests to Pull Directory Servers are typically derived from
ingressed ARP [RFC826], ND [RFC4861], or RARP [RFC903] messages, or
data frames with unknown unicast destination MAC addresses,
intercepted by an ingress TRILL switch as described in Section 4.
Pull Directory responses include an amount of time for which the
response should be considered valid. This includes negative responses
that indicate no data is available. Thus both positive responses with
data and negative responses can be cached and used to locally handle
ARP, ND, RARP, unknown destination MAC frames, or the like, until the
responses expire. If information previously pulled is about to
expire, a TRILL switch MAY try to refresh it by issuing a new pull
request but, to avoid unnecessary requests, SHOULD NOT do so if it
has not been recently used. The validity timer of cached Pull
Directory responses is NOT reset or extended merely because that
cache entry is used.
L. Dunbar, et al [Page 12]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
3.1 Pull Directory Message Common Format
All Pull Directory messages are transmitted as the payload of RBridge
Channel messages. All Pull Directory messages are formatted as
described below starting with the following common 8-byte header:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Ver | Type | Flags | Count | Err | SubErr |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Sequence Number |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Type Specific Payload - variable length
+-+-+- ...
Ver: Version of the Pull Directory protocol as an unsigned
integer. Version zero is specified in this document.
Type: The Pull Directory message type as follows:
Type Section Name
---- ------- --------
0 3.2.1 Query
1 3.2.2 Response
2 3.1.4 Update
3 3.1.5 Acknowledge
4-15 - Reserved
Flags: Four flag bits whose meaning depends on the Pull Directory
message Type. Flags whose meaning is not specified are
reserved, MUST be sent as zero, and MUST be ignored on receipt.
Count: Most Pull Directory message types specified herein have
zero or more occurrences of a Record as part of the type
specific payload. The Count field is the number of occurrences
of that Record as an unsigned integer. For Pull Directory
messages not structured with such occurrences, this field MUST
be sent as zero and ignored on receipt.
Err, SubErr: The error and suberror fields are only used in
messages that are in the nature of replies or acknowledgements.
In messages that are requests or updates, these fields MUST be
sent as zero and ignored on receipt. The meaning of values in
the Err field depends on the Pull Directory message Type but in
all cases the value zero means no error. The meaning of values
in the SubErr field depends on both the message Type and on the
value of the Err field but in all cases, a zero SubErr field is
allowed and provides no additional information beyond the value
of the Err field.
L. Dunbar, et al [Page 13]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Sequence Number: An opaque 32-bit quantity set by the TRILL switch
sending a request or other unsolicited message and returned in
every corresponding reply or acknowledgement. It is used to
match up responses with the message to which they respond.
Type Specific Payload: Format depends on the Pull Directory
message Type.
3.2 Pull Directory Query and Response Messages
3.2.1 Pull Directory Query Message Format
A Pull Directory Query message is sent as the Channel Protocol
specific content of an RBridge Channel message [RFC7178] TRILL Data
packet or as a native RBridge Channel data frame (see Section 3.4).
The Data Label of the packet is the Data Label in which the query is
being made. The priority of the channel message is a mapping of the
priority of the frame being ingressed that caused the query with the
default mapping depending, per Data Label, on the strategy (see
Section 4) or a configured priority for generated queries. (Geerate
queries are those not the result of a mapping. For example, a query
to refresh a cache entry.) The Channel Protocol specific data is
formatted as a header and a sequence of zero or more QUERY Records as
follows:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Ver | Type | Flags | Count | Err | SubErr |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Sequence Number |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| QUERY 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| QUERY 2
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| QUERY K
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
Ver, Sequence Number: See 3.1.
Type: 1 for Query. Queries received by an TRILL switch that is not
a Pull Directory result in an error response (see Section 3.5)
unless inhibited by rate limiting.
L. Dunbar, et al [Page 14]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Flags, Err, and SubErr: MUST be sent as zero and ignored on
receipt.
Count: Number of QUERY Records present. A Query message Count of
zero is explicitly allowed, for the purpose of pinging a Pull
Directory server to see if it is responding. On receipt of such
an empty Query message, a Response message that also has a
Count of zero is sent unless inhibited by rate limiting.
QUERY: Each QUERY Record within a Pull Directory Query message is
formatted as follows:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| SIZE | RESV | QTYPE |
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
If QTYPE = 1
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| AFN |
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| Query address ...
+--+--+--+--+--+--+--+--+--+--+--...
If QTYPE = 2, 3, 4, or 5
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| Query frame ...
+--+--+--+--+--+--+--+--+--+--+--...
SIZE: Size of the QUERY record in bytes as an unsigned integer
starting not counting the SIZE field and following byte.
Thus the minimum legal value is 2. A value of SIZE less than
2 indicates a malformed QUERY record. The QUERY record with
the illegal SIZE value and any subsequent QUERY records MUST
be ignored and the entire Query message MAY be ignored.
RESV: A block of reserved bits. MUST be sent as zero and
ignored on receipt.
QTYPE: There are several types of QUERY Records currently
defined in two classes as follows: (1) a QUERY Record that
provides an explicit address and asks for all addresses for
the interface specified by the query address and (2) a QUERY
Record that includes a frame. The fields of each are
specified below. Values of QTYPE are as follows:
L. Dunbar, et al [Page 15]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
QTYPE Description
----- -----------
0 reserved
1 address query
2 ARP query frame
3 ND query frame
4 RARP query frame
5 Unknown unicast MAC query frame
6-14 assignable by IETF Review
15 reserved
AFN: Address Family Number of the query address.
Address Query: The query is asking for any other addresses,
and the nickname of the TRILL switch from which they are
reachable, that correspond to the same interface, within
the data label of the query. Typically that would be
either (1) a MAC address with the querying TRILL switch
primarily interested in the TRILL switch by which that
MAC address is reachable, or (2) an IP address with the
querying TRILL switch interested in the corresponding MAC
address and the TRILL switch by which that MAC address is
reachable. But it could be some other address type.
Query Frame: Where a QUERY Record is the result of an ARP,
ND, RARP, or unknown unicast MAC destination address, the
ingress TRILL switch MAY send the frame to a Pull
Directory Server if the frame is small enough that the
resulting Query message fits into a TRILL Data packet
within the campus MTU.
If no response is received to a Pull Directory Query message within a
timeout configurable in milliseconds that defaults to 200, the Query
message should be re-transmitted with the same Sequence Number up to
a configurable number of times that defaults to three. If there are
multiple QUERY Records in a Query message, responses can be received
to various subsets of these QUERY Records before the timeout. In that
case, the remaining unanswered QUERY Records should be re-sent in a
new Query message with a new sequence number. If a TRILL switch is
not capable of handling partial responses to queries with multiple
QUERY Records, it MUST NOT sent a Request message with more than one
QUERY Record in it.
See Section 3.5 for a discussion of how Query message errors are
handled.
L. Dunbar, et al [Page 16]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
3.2.2 Pull Directory Response Format
Pull Directory Response messages are sent as the Channel Protocol
specific content of an RBridge Channel message [RFC7178] TRILL Data
packet or as a native RBridge Channel data frame (see Section 3.4).
Responses are sent with the same Data Label and priority as the Query
message to which they correspond except that the Response message
priority is limited to be not more than a configured value. This
priority limit is configurable at per TRILL switch and defaults to
priority 6. Pull Directory Response messages SHOULD NOT be sent with
priority 7 as that priority SHOULD be reserved for messages critical
to network connectivity.
The RBridge Channel protocol specific data format is as follows:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Ver | Type | Flags | Count | Err | SubErr |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Sequence Number |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| RESPONSE 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| RESPONSE 2
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| RESPONSE K
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
Ver, Sequence Number: As specified in Section 3.1.
Type: 2 = Response.
Flags: MUST be sent as zero and ignored on receipt.
Count: Count is the number of RESPONSE Records present in the
Response message.
Err, SubErr: A two part error code. Zero unless there was an error
in the Query message, for which case see Section 3.5.
RESPONSE: Each RESPONSE record within a Pull Directory Response
message is formatted as follows:
L. Dunbar, et al [Page 17]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| SIZE |OV| RESV | Index |
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| Lifetime |
+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
| Response Data ...
+--+--+--+--+--+--+--+--+--+--+--...
SIZE: Size of the RESPONSE Record in bytes not counting the
SIZE field and following byte. Thus the minimum value of
SIZE is 2. If SIZE is less than 2, that RESPONSE Record and
all subsequent RESPONSE Records in the Response message MUST
be ignored and the entire Response message MAY be ignored.
OV: The overflow flag. Indicates, as described below, that
there was too much Response Data to include in one Response
message.
RESV: Three reserved bits that MUST be sent as zero and ignored
on receipt.
Index: The relative index of the QUERY Record in the Query
message to which this RESPONSE Record corresponds. The index
will always be one for Query messages containing a single
QUERY Record. If the Index is larger than the Count was in
the corresponding Query, that RESPONSE Record MUST be
ignored and subsequent RESPONSE Records or the entire
Response message MAY be ignored.
Lifetime: The length of time for which the response should be
considered valid in units of 200 milliseconds except that
the values zero and 2**16-1 are special. If zero, the
response can only be used for the particular query from
which it resulted and MUST NOT be cached. If 2**16-1, the
response MAY be kept indefinitely but not after the Pull
Directory server goes down or becomes unreachable. The
maximum definite time that can be expressed is a little over
3.6 hours.
Response Data: There are various types of RESPONSE Records.
- If the Err field is non-zero, then the Response Data is a
copy of the corresponding QUERY Record data, that is,
either an AFN followed by an address or a query frame.
See Section 3.5 for additional information on errors.
- If the Err field is zero and the corresponding QUERY
Record was an address query, then the Response Data is
formated as the value of an Interface Addresses APPsub-
TLV [IA]. The maximum size of such contents is 253 bytes
in the case when SIZE is 255.
L. Dunbar, et al [Page 18]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
- If the Err field is zero and the corresponding QUERY
Record was a frame query, then the Response data consists
of the response frame for ARP, ND, or RARP and a copy of
the frame for unknown unicast destination MAC.
Multiple RESPONSE Records can appear in a Response message with the
same index if the answer to a QUERY Record consists of multiple
Interface Address APPsub-TLV values. This would be necessary if, for
example, a MAC address within a Data Label appears to be reachable by
multiple TRILL switches. However, all RESPONSE Records to any
particular QUERY Record MUST occur in the same Response message. If a
Pull Directory holds more mappings for a queried address than will
fit into one Response message, it selects which to include by some
method outside the scope of this document and sets the overflow flag
(OV) in all of the RESPONSE Records responding to that query address.
See Section 3.5 for a discussion of how errors are handled.
3.3 Cache Consistency
A Pull Directory MUST take action to minimize the amount of time that
a TRILL switch will continue to use stale information from that Pull
Directory by sending Update messages.
A Pull Directory server MUST maintain one of the following three sets
of records, in order of increasing specificity. Retaining more
specific records, such as that given in item 3 below, minimizes
Spontaneous Update messages sent to update pull client TRILL switch
caches but increases the record keeping burden on the Pull Directory
server. Retaining less specific records, such as that given in item
1, will generally increase the volume and overhead due to Spontaneous
Update messages and due to unnecessarily invalidating cached
information, but will still maintain consistency and will reduce the
record keeping burden on the Pull Directory server. In all cases,
there may still be brief periods of time when directory information
has changed but cached information a pull clients has not yet been
updated or expunged.
1. An overall record per Data Label of when the last positive
response data sent will expire at some requester and when the
last negative response will expire at some requester, assuming
those responders cached the response.
2. For each unit of data (IA APPsub-TLV Address Set [IA]) held by
the server and each address about which `a negative response
was sent, when the last response sent with that positive
response data or negative response will expire at a requester,
assuming the requester cached the response.
L. Dunbar, et al [Page 19]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
3. For each unit of data held by the server (IA APPsub-TLV Address
Set [IA]) and each address about which a negative response was
sent, a list of TRILL switches that were sent that data as a
positive response or sent a negative response for the address,
and the expected time to expiration for that data or address at
each such TRILL switch, assuming the requester cached the
response.
A Pull Directory server may have a limit as to how many TRILL
switches for which it can maintain expiry information by method 3
above or how many data units or addresses it can maintain expiry
information for by method 2. If such limits are exceeded, it MUST
transition to a lower numbered strategy but, in all cases, MUST
support, at a minimum, method 1.
When data at a Pull Directory changes or is deleted or data is added
and there may be unexpired stale information at a requesting TRILL
switch, the Pull Directory MUST send an Update message as discussed
below. The sending of such an Update message MAY be delayed by a
configurable number of milliseconds that default to 50 milliseconds
to await other possible changes that could be included in the same
Update.
If method 1, the most crude method, is being followed, then when any
Pull Directory information in a Data Label is changed or deleted and
there are outstanding cached positive data response(s), an all-
addresses flush positive Update message is flooded within that Data
Label as an RBridge Channel message with an Inner.MacDA of All-
Egress-RBridges. And if data is added and there are outstanding
cached negative responses, an all-addresses flush negative message is
similarly flooded. "All-addresses" is indicated by the Count field
being zero in an Update message. On receiving an all-addresses
flooded flush positive Update from a Pull Directory server it has
used, indicated by the F and P bits being one and the Count being
zero, a TRILL switch discards all cached data responses it has for
that Data Label. Similarly, on receiving an all addresses flush
negative Update, indicated by the F and N bits being one and the
Count being zero, it discards all cached negative replies for that
Data Label. A combined flush positive and negative can be flooded by
having all of the F, P, and N bits set to one resulting in the
discard of all positive and negative cached information for the Data
Label.
If method 2 is being followed, then a TRILL switch floods address
specific positive Update messages when data that might be cached by a
querying TRILL switch is changed or deleted and floods address
specific negative Update messages when such information is added to.
Such messages are similar to the method 1 flooded flush Update
messages and are also sent as RBridge Channel messages with an
Inner.MacDA of All-Egress-RBridges. However the Count field will be
L. Dunbar, et al [Page 20]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
non-zero and either the P or N bit, but not both, will be one. On
receiving such as address specific unsolicited update, if it is
positive the addresses in the RESPONSE records in the unsolicited
response are compared to the addresses about which the receiving
TRILL switch is holding cached positive information from that server
and, if they match, the cached information is updated. On receiving
an address specific unsolicited update negative message, the
addresses in the RESPONSE records in the unsolicited update are
compared to the addresses about which the receiving TRILL switch is
holding cached negative information from that server and, if they
match, the cached negative information is updated.
If method 3 is being followed, the same sort of unsolicited update
messages are sent as with method 2 above except they are not normally
flooded but unicast only to the specific TRILL switches the directory
server believes may be holding the cached positive or negative
information that needs updating. However, a Pull Directory server MAY
flood the unsolicited update under method 3, for example if it
determines that a sufficiently large fraction of the TRILL switches
in some Data label are requesters that need to be updated.
A Pull Directory server tracking cached information with method 3
MUST NOT clear the indication that it needs update cached information
at a querying TRILL switch until it has sent an Update message and
received a corresponding Acknowledge message or it has sent a
configurable number of updates at a configurable interval which
default to 3 updates 200 milliseconds apart.
A Pull Directory server tracking cached information with methods 2 or
1 SHOULD NOT clear the indication that it needs to update cached
information until it has sent an Update message and received a
corresponding Acknowledge message from all of its ESADI neighbors or
it has sent a configurable number of updates at a configurable
interval that defaults to 3 updates 200 milliseconds apart.
3.3.1 Update Message Format
An Update message is formatted as a Response message except that the
Type field in the message header is a different value.
Update messages are initiated by a Pull Directory server. The
Sequence number space used is controlled by the originating Pull
Directory server and different from Sequence number space used in a
Query and the corresponding Response that are controlled by the
querying TRILL switch.
The Flags field of the message header for an Update message is as
follows:
L. Dunbar, et al [Page 21]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
+---+---+---+---+
| F | P | N | R |
+---+---+---+---+
F: The Flood bit. If zero, the response is to be unicast . If F=1, it
is multicast to All-Egress-RBridges.
P, N: Flags used to indicate positive or negative Update messages.
P=1 indicates positive. N=1 indicates negative. Both may be 1 for
a flooded all addresses Update.
R: Reserved. MUST be sent as zero and ignored on receipt
3.3.2 Acknowledge Message Format
An Acknowledge message is sent in response to an Update to confirm
receipt or indicate an error unless response is inhibited by rate
limiting. It is also formatted as a Response message.
If there are no errors in the processing of an Update message, the
message is essentially echoed back with the Type changed to
Acknowledge.
If there was an overall or header error in an Update message, it is
echoed back as an Acknowledge message with the Err and SubErr fields
set appropriately (see Section 3.5).
If there is a RESPONSE Record level error in an Update message, one
or more Acknowledge messages may be returns as indicated in Section
3.5.
3.4 Pull Directory Hosted on an End Station
Optionally, a Pull Directory actually hosted on an end station MAY be
supported. In that case, one or more TRILL switches must proxy for
the end station and advertise themselves as a Pull Directory server.
Such proxies must have a direct connection to the end station, that
is a connection not involving any intermediate TRILL switches.
When the proxy TRILL switch receives a Query message, it modifies the
inter-RBridge Channel message received into a native RBridge Channel
message and forwards it to that end station. Later, when it receives
one or more responses from that end station by native RBridge Channel
messages, it modifies them into inter-RBridge Channel messages and
forwards them to the source TRILL switch of the original Query
message. Similarly, an Update from the end station is forwarded to
L. Dunbar, et al [Page 22]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
client TRILL switches and acknowledgements from those TRILL switches
are returned to the end station by the proxy. Because native RBridge
Channel messages have no TRILL Header and are addressed by MAC
address, as opposed to inter-RBridge Channel messages that are TRILL
Data packets and are addressed by nickname, nickname information must
be added to the native RBridge Channel version of Pull Directory
messages.
The native Pull Directory RBridge Channel messages use the same
Channel protocol number as do the inter-RBridge Pull Directory
RBridge Channel messages. The native messages SHOULD be sent with an
Outer.VLAN tag which gives the priority of each message which is the
priority of the original inter-RBridge request packet. The Outer.VLAN
ID used is the Designated VLAN on the link to the end station. Since
there is no TRILL Header or inner Data Label for native RBridge
Chanel messages, that information is added to the header.
The native RBridge Channel message Pull Directory message protocol
dependent data part is the same as for inter-RBridge Channel messages
except that the 8-byte header described in Section 3.1 is expanded to
14 or 18 bytes as follows:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Ver | Type | Flags | Count | Err | SubErr |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Sequence Number |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Nickname (2 bytes) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+...+-+
| Data Label ... (4 or 8 bytes) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+...+-+
| Type Specific Payload - variable length
+-+-+- ...
Fields not described below are as in Section 3.1.
Data Label: The Data Label that normally appear right after the
Inner.MacSA of the an RBridge Channel Pull Directory message
appears here in the native RBridge Channel message version.
This might appear in a Query message, to be reflected in a
Response message, or it might appear in an Update message, to
be reflected in an Acknowledge message.
Nickname: The nickname of the TRILL switch that is communicating
with the end station Pull Directory. Usually this is a remote
TRILL switch but it could be the TRILL switch to which the end
station is attached. The proxy copies this from the ingress
nickname when mapping a Query or Acknowledge message to native
L. Dunbar, et al [Page 23]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
form. It also takes this from a native Response or Update to be
used as the egress of the inter-RBridge form on the message
unless it is a flooded Update in which case a distribution tree
is used.
3.5 Pull Directory Message Errors
A non-zero Err field in the Pull Directory message header indicates
an error message.
If there is an error that applies to an entire Query message or its
header, as indicated by the range of the value of the Err field, then
the QUERY records in the request are just echoed back in the RESPONSE
records of the Response message but expanded with a zero Lifetime and
the insertion of the Index field. If there is an error that applies
to an entire Update message or its header, then the RESPONSE records
in the update, if any, are echoed back in the Acknowledge message.
If errors occur at the QUERY Record level for a Query message, they
MUST be reported in a Response message separate from the results of
any successful non-erroneous QUERY Records. If multiple QUERY Records
in a Query message have different errors, they MUST be reported in
separate Response messages. If multiple QUERY Records in a Query
message have the same error, this error response MAY be reported in
one or multiple Response messages. In an error Response message, the
QUERY Record or records being responded to appear, expanded by the
Lifetime for which the server thinks the error might persist and with
their Index inserted, as the RESPONSE record or records.
If errors occur at the RESPONSE Record level for an Update message,
they MUST be reported in a Acknowledge message separate from the
acknowledgement of any non-erroneous RESPONSE Records. If multiple
RESPONSE Records in an Update have different errors, they MUST be
reported in separate Acknowledge messages. If multiple RESPONSE
Records in an Update message have the same error, this error response
MAY be reported in one or multiple Acknowledge messages. In an error
Acknowledge message, the RESPONSE Record or records being responded
to appear, expanded by the time for which the server thinks the error
might persist and with their Index inserted, as a RESPONSE Record or
records.
ERR values 1 through 127 are available for encoding Request or Update
message level errors. ERR values 128 through 254 are available for
encoding QUERY or RESPONSE Record level errors. The SubErr field is
available for providing more detail on errors. The meaning of a
SubErr field value depends on the value of the Err field.
L. Dunbar, et al [Page 24]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Err Meaning
--- -------
0 (no error)
1 Unknown or reserved Query message field value
2 Request data too short
3 Unknown or reserved Update message field value
4 Update data too short
5-127 (Available for allocation by IETF Review)
128 Unknown or reserved QUERY Record field value
129 Address not found
130 Unknown or reserved RESPONSE Record field value
131-254 (Available for allocation by IETF Review)
255 Reserved
The following sub-errors are specified under error code 1 and 3:
SubErr Field with Error
------ ----------------
0 Unspecified
1 Unknown V field value
2 Reserved T field value
3 Zero sequence number in request
4-254 (Available for allocation by Expert Review)
255 Reserved
The following sub-errors are specified under error code 128 and 130:
SubErr Field with Error
------ ----------------
0 Unspecified
1 Unknown AFN field value
2 Unknown or Reserved TYPE field value
3 Invalid or inconsistent SIZE field value
4-254 (Available for allocation by Expert Review)
255 Reserved
More TBD
3.6 Additional Pull Details
If a TRILL switch notices that a Pull Directory server is no longer
data reachable [RFC7180], it MUST promptly discard all pull responses
it is retaining from that server as it can no longer receive cache
L. Dunbar, et al [Page 25]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
consistency update messages from the server.
Because a Pull Directory server may need to advertise interest in
Data Labels even though it does not want to received end station data
in those Data Labels, the No Data (NOD) flag bit is provided as
specified in Section 6.3. For example, an RBridge hosting a Pull
Directory may be a secondary directory that wants to receive its data
from a primary Push Directory server but have no interest in
receiving multicast traffic from end stations.
L. Dunbar, et al [Page 26]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
4. Directory Use Strategies and Push-Pull Hybrids
For some edge nodes that have a great number of Data Labels enabled,
managing the MAC and Data Label <-> Edge RBridge mapping for hosts
under all those Data Labels can be a challenge. This is especially
true for Data Center gateway nodes, which need to communicate with a
majority of Data Labels, if not all.
For those edge TRILL switch nodes, a hybrid model should be
considered. That is the Push Model is used for some Data Labels, and
the Pull Model is used for other Data Labels. It is the network
operator's decision by configuration as to which Data Labels' mapping
entries are pushed down from directories and which Data Labels'
mapping entries are pulled.
For example, assume a data center where hosts in specific Data
Labels, say VLANs 1 through 100, communicate regularly with external
peers. Probably, the mapping entries for those 100 VLANs should be
pushed down to the data center gateway routers. For hosts in other
Data Labels which only communicate with external peers occasionally
for management interface, the mapping entries for those VLANs should
be pulled down from directory when the need comes up.
The mechanisms described above for Push and Pull Directory services
make it easy to use Push for some Data Labels and Pull for others. In
fact, different TRILL switches can even be configured so that some
use Push Directory services and some use Pull Directory services for
the same Data Label if both Push and Pull Directory services are
available for that Data Label. And there can be Data Labels for which
directory services are not used at all.
For Data Labels in which a hybrid push/pull approach is being taken,
it would make sense to use push for address information of hosts that
frequently communicate with many other hosts in the Data Label, such
as a file or DNS server. Pull could then be used for hosts that
communicate with few other hosts, perhaps such as hosts being used as
compute engines.
4.1 Strategy Configuration
Each TRILL switch that has the ability to use directory assistance
has, for each Data Label X in which it is might ingress native
frames, one of four major modes:
0. No directory use: The TRILL switch does not subscribe to Push
Directory data or make Pull Directory requests for Data Label X
and directory data is not consulted on ingressed frames in Data
Label X that might have used directory data. This includes ARP,
L. Dunbar, et al [Page 27]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
ND, RARP, and unknown MAC destination addresses, which are
flooded as appropriate.
1. Use Push only: The TRILL switch subscribes to Push Directory
data for Data Label X.
2. Use Pull only: When the TRILL switch ingresses a frame in Data
Label X that can use Directory information, if it has cached
information for the address it uses it. If it does not have
either cached positive or negative information for the address,
it sends a Pull Directory query.
3. Use Push and Pull: The TRILL switch subscribes to Push
Directory data for Data Label X. When it ingresses a frame in
Data Label X that can use Directory information and it does not
find that information in its link state database of Push
Directory information, it makes a Pull Directory query.
The above major Directory use mode is per Data Label. In addition,
there is a per Data Label per priority minor mode as listed below
that indicates what should be done if Directory Data is not available
for the ingressed frame. In all cases, if you are holding Push
Directory or Pull Directory information to handle the frame given the
major mode, the directory information is simply used and, in that
instance, the minor mode does not matter.
A. Flood immediate: Flood the frame immediately (even if you are
also sending a Pull Directory) request.
B. Flood: Flood the frame immediately unless you are going to do a
Pull Directory request, in which case you wait for the response
or for the request to time out after retries and flood the
frame if the request times out.
C. Discard if complete or Flood immediate: If you have complete
Push Directory information and the address is not in that
information, discard the frame. If you do not have complete
Push Directory information, the same as A above.
D. Discard if complete or Flood: If you have complete Push
Directory information and the address is not in that
information, discard the frame. If you do not have complete
Push Directory information, the same as B above.
In addition, the query message priority for Pull Directory requests
sent can be configured on a per Data Label, per ingressed frame
priority basis. The default mappings are as follows where Ingress
Priority is the priority of the native frame that provoked the Pull
Directory query:
L. Dunbar, et al [Page 28]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Ingress If Flood If Flood
Priority Immediate Delayed
-------- --------- --------
7 5 6
6 5 6
5 4 5
4 3 4
3 2 3
2 0 2
0 1 0
1 1 1
Priority 7 is normally only used for urgent messages critical to
adjacency and so is avoided by default for directory traffic.
Unsolicited updates are sent with a priority that is configured per
Data Label that defaults to priority 5.
L. Dunbar, et al [Page 29]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
5. Security Considerations
Incorrect directory information can result in a variety of security
threats including the following:
Incorrect directory mappings can result in data being delivered to
the wrong end stations, or set of end stations in the case of
multi-destination packets, violation security policy.
Missing or incorrect directory data can result in denial of
service due to sending data packets to black holes or discarding
data on ingress due to incorrect information that their
destinations are not reachable.
Push Directory data is distributed through ESADI-LSPs [RFC7357] that
can be authenticated with the same mechanisms as IS-IS LSPs. See
[RFC5304] [RFC5310] and the Security Considerations section of
[RFC7357].
Pull Directory queries and responses are transmitted as RBridge-to-
RBridge or native RBridge Channel messages. Such messages can be
secured as specified in [ChannelTunnel].
For general TRILL security considerations, see [RFC6325].
L. Dunbar, et al [Page 30]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
6. IANA Considerations
This section gives IANA assignment and registry considerations.
6.1 ESADI-Parameter Data Extensions
IANA will assigned two ESADI-Parameter TRILL APPsub-TLV flag bits for
"Push Directory" (PSH) and "Complete Push" (COP) and will create a
sub-registry in the TRILL Parameters Registry as follows:
Sub-Registry: ESADI-Parameter APPsub-TLV Flag Bits
Registration Procedures: Standards Action
References: [RFC7357] [This document]
Bit Mnemonic Description Reference
--- -------- ----------- ---------
0 UN Supports Unicast ESADI ESDADI [RFC7357]
1 PSH Push Directory Server This document
2 COP Complete Push This document
3-7 - available for allocation
The COP bit is ignored if the PSH bit is zero.
In addition, the ESADI-Parameter APPsub-TLV is optionally extended,
as provided in its original specification in ESDADI [RFC7357], by one
byte as show below:
+-+-+-+-+-+-+-+-+
| Type | (1 byte)
+-+-+-+-+-+-+-+-+
| Length | (1 byte)
+-+-+-+-+-+-+-+-+
|R| Priority | (1 byte)
+-+-+-+-+-+-+-+-+
| CSNP Time | (1 byte)
+-+-+-+-+-+-+-+-+
| Flags | (1 byte)
+---------------+
|PushDirPriority| (optional, 1 byte)
+---------------+
| Reserved for expansion (variable)
+-+-+-+-...
The meanings of all the fields are as specified in ESDADI [RFC7357]
except that the added PushDirPriority is the priority of the
advertising ESADI instance to be a Push Directory as described in
L. Dunbar, et al [Page 31]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Section 2.3. If the PushDirPriority field is not present (Length = 3)
it is treated as if it were 0x40. 0x40 is also the value used and
placed here by an TRILL switch whose priority to be a Push Directory
has not been configured.
6.2 RBridge Channel Protocol Number
IANA will allocate a new RBridge Channel protocol number for "Pull
Directory Services" from the range allocable by Standards Action and
update the subregistry of such protocol number in the TRILL
Parameters Registry referencing this document.
6.3 The Pull Directory (PUL) and No Data (NOD) Bits
IANA is requested to allocate two currently reserved bits in the
Interested VLANs field of the Interested VLANs sub-TLV (suggested
bits 18 and 19) and the Interested Labels field of the Interested
Labels sub-TLV (suggested bits 6 and 7) [RFC7176] to indicate Pull
Directory server (PUL) and No Data (NOD) respectively. These bits are
to be added, with this document as reference, to the "Interested
VLANs Flag Bits" and "Interested Labels Flag Bits" subregistries
created by [RFC7357].
{{Material below in this subsection is technical and should be moved
out of the IANA Consdierations.}}
In the TRILL base protocol [RFC6325] as extended for FGL [RFC7172],
the mere presence of an Interested VLANs or Interested Labels sub-
TLVs in the LSP of a TRILL switch indicates connection to end
stations in the VLAN(s) or FGL(s) listed and thus a desire to receive
multi-destination traffic in those Data Labels. But, with Push and
Pull Directories, advertising that you are a directory server
requires using these sub-TLVs to indicate the Data Label(s) you are
serving. If such a directory server does not wish to received multi-
destination TRILL Data packets for the Data Labels it lists in one of
these sub-TLVs, it sets the "No Data" (NOD) bit to one. This means
that data on a distribution tree may be pruned so as not to reach the
"No Data" TRILL switch as long as there are no TRILL switches
interested in the Data that are beyond the "No Data" TRILL switch on
a distribution tree. The NOD bit is backwards compatible as TRILL
switches ignorant of it will simply not prune when they could, which
is safe although it may cause increased link utilization.
Example of a TRILL switch serving as a directory that might not want
multi-destination traffic in some Data Labels would be a TRILL switch
that does not offer end station service for any of the Data Labels
L. Dunbar, et al [Page 32]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
for which it is serving as a directory and is either
- a Pull Directory and/or
- a Push Directory for which all of the ESADI traffic will be
handled by unicast ESDADI [RFC7357].
A Push Directory MUST NOT set the NOD bit for a data label if it
needs to communicate via multi-destination ESADI PDUs in that data
label since such PDUs look like TRILL Data packets to transit TRILL
switches and might be incorrectly pruned if NOD was set.
L. Dunbar, et al [Page 33]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Acknowledgments
The contributions of the following persons are gratefully
acknowledged:
TBD
The document was prepared in raw nroff. All macros used were defined
within the source file.
L. Dunbar, et al [Page 34]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Normative References
[RFC826] - Plummer, D., "An Ethernet Address Resolution Protocol",
RFC 826, November 1982.
[RFC903] - Finlayson, R., Mann, T., Mogul, J., and M. Theimer, "A
Reverse Address Resolution Protocol", STD 38, RFC 903, June
1984
[RFC2119] - Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997
[RFC3971] - Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander,
"SEcure Neighbor Discovery (SEND)", RFC 3971, March 2005.
[RFC4861] - Narten, T., Nordmark, E., Simpson, W., and H. Soliman,
"Neighbor Discovery for IP version 6 (IPv6)", RFC 4861,
September 2007.
[RFC5304] Li, T. and R. Atkinson, "IS-IS Cryptographic
Authentication", RFC 5304, October 2008.
[RFC5310] - Bhatia, M., Manral, V., Li, T., Atkinson, R., White, R.,
and M. Fanto, "IS-IS Generic Cryptographic Authentication", RFC
5310, February 2009.
[RFC6165] - Banerjee, A. and D. Ward, "Extensions to IS-IS for
Layer-2 Systems", RFC 6165, April 2011.
[RFC6325] - Perlman, R., Eastlake 3rd, D., Dutt, D., Gai, S., and A.
Ghanwani, "Routing Bridges (RBridges): Base Protocol
Specification", RFC 6325, July 2011.
[RFC7042] - Eastlake 3rd, D. and J. Abley, "IANA Considerations and
IETF Protocol and Documentation Usage for IEEE 802 Parameters",
BCP 141, RFC 7042, October 2013.
[RFC7172] - Eastlake 3rd, D., Zhang, M., Agarwal, P., Perlman, R.,
and D. Dutt, "Transparent Interconnection of Lots of Links
(TRILL): Fine-Grained Labeling", RFC 7172, May 2014,
<http://www.rfc-editor.org/info/rfc7172>.
[RFC7176] - Eastlake 3rd, D., Senevirathne, T., Ghanwani, A., Dutt,
D., and A. Banerjee, "Transparent Interconnection of Lots of
Links (TRILL) Use of IS-IS", RFC 7176, May 2014,
<http://www.rfc-editor.org/info/rfc7176>.
[RFC7178] - Eastlake 3rd, D., Manral, V., Li, Y., Aldrin, S., and D.
Ward, "Transparent Interconnection of Lots of Links (TRILL):
RBridge Channel Support", RFC 7178, May 2014, <http://www.rfc-
L. Dunbar, et al [Page 35]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
editor.org/info/rfc7178>.
[RFC7180] - Eastlake 3rd, D., Zhang, M., Ghanwani, A., Manral, V.,
and A. Banerjee, "Transparent Interconnection of Lots of Links
(TRILL): Clarifications, Corrections, and Updates", RFC 7180,
May 2014, <http://www.rfc-editor.org/info/rfc7180>.
[RFC7357] - Zhai, H., Hu, F., Perlman, R., Eastlake 3rd, D., and O.
Stokes, "Transparent Interconnection of Lots of Links (TRILL):
End Station Address Distribution Information (ESADI) Protocol",
RFC 7357, September 2014, <http://www.rfc-
editor.org/info/rfc7357>.
[IA] - Eastlake, D., L. Yizhou, R. Perlman, "TRILL: Interface
Addresses APPsub-TLV", draft-eastlake-trill-ia-appsubtlv, work
in progress.
Informational References
[RFC7067] - Dunbar, L., Eastlake 3rd, D., Perlman, R., and I.
Gashinsky, "Directory Assistance Problem and High-Level Design
Proposal", RFC 7067, November 2013.
[ChannelTunnel] - D. Eastlake, Y. Li, "TRILL: RBridge Channel Tunnel
Protocol", draft-eastlake-trill-channel-tunnel, work in
progress.
[ARP reduction] - Shah, et. al., "ARP Broadcast Reduction for Large
Data Centers", Oct 2010.
L. Dunbar, et al [Page 36]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Authors' Addresses
Linda Dunbar
Huawei Technologies
5430 Legacy Drive, Suite #175
Plano, TX 75024, USA
Phone: +1-469-277-5840
Email: ldunbar@huawei.com
Donald Eastlake
Huawei Technologies
155 Beaver Street
Milford, MA 01757 USA
Phone: +1-508-333-2270
Email: d3e3e3@gmail.com
Radia Perlman
EMC
2010 256th Avenue NE, #200
Bellevue, WA 98007 USA
Email: Radia@alum.mit.edu
Igor Gashinsky
Yahoo
45 West 18th Street 6th floor
New York, NY 10011
Email: igor@yahoo-inc.com
Yizhou Li
Huawei Technologies
101 Software Avenue,
Nanjing 210012 China
Phone: +86-25-56622310
Email: liyizhou@huawei.com
L. Dunbar, et al [Page 37]
INTERNET-DRAFT TRILL: Directory Service Mechanisms
Copyright, Disclaimer, and Additional IPR Provisions
Copyright (c) 2014 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License. The definitive version of
an IETF Document is that published by, or under the auspices of, the
IETF. Versions of IETF Documents that are published by third parties,
including those that are translated into other languages, should not
be considered to be definitive versions of IETF Documents. The
definitive version of these Legal Provisions is that published by, or
under the auspices of, the IETF. Versions of these Legal Provisions
that are published by third parties, including those that are
translated into other languages, should not be considered to be
definitive versions of these Legal Provisions. For the avoidance of
doubt, each Contributor to the IETF Standards Process licenses each
Contribution that he or she makes as part of the IETF Standards
Process to the IETF Trust pursuant to the provisions of RFC 5378. No
language to the contrary, or terms, conditions or rights that differ
from or are inconsistent with the rights and licenses granted under
RFC 5378, shall have any effect and shall be null and void, whether
published or posted by such Contributor, or included with or in such
Contribution.
L. Dunbar, et al [Page 38]
