Watch, Fabric Access, and the BROCADE logo are t rademarks or registered trademarks of Brocade Communications
Systems, Inc., in the United States and/or in other countries.
All other brands, products, or service names are or may be trademarks or service marks of, and are used to identify,
products or services of their respect ive owners.
ROCADE. BROCADERESERVESTHERIGHTTOMAKECHANGESTOTHISDOCUMENTATANYTIME, W ITHOUTNOTICE, AND
B
ASSUMESNORESPONSIBILITYFORITSUSE.
Export of technical data contained in this document may require an export license from the U ni ted States Government.
Brocade Communications Systems, Incorporated
Fabric Watch is an optionally licensed product, and requires a valid license key to function. It is
supported on the SilkWorm® 3800 series of switches running Brocades Fabric OS™ v3.0.
About This Guide
This guide provides the following information about Fabric Watch:
Chapter 1
Introducing Fabric Watch
Chapter 2
Installing Fabric Watch
Chapter 3
Using Fabric Watch
Related Publications
Related product information can be found in the following Brocade publications:
•Brocade Fabric OS Procedures Guide
•Brocade Fabric OS Reference
•Brocade MIB Reference
•Brocade Web Tools User’s Guide
•Brocade Zoning User’s Guide
•Brocade QuickLoop User’s Guide
•Brocade Fabric Watch User’s Guide
Provides an overview of Fabric Watch.
Provides instructions for installing Fabric Watch.
Provides information about configuring and using Fabric
Watch.
Information about fibre channel standards and fibre channel in general can be found on the Fibre
Channel Association web site, located at:
http://www.fibrechannel.com
Getting Help
Contact your switch supplier for technical support. This includes hardware and software support,
all product repairs, and ordering of spare components.
Be prepared to provide the following information to support personnel:
•Switch serial number
•Switch worldwide name
•Topology configuration
•Output from the supportShow telnet command
•Detailed description of the problem
•Troubleshooting steps already performed
Getting Software Updates
Contact your switch supplier for software updates and maintenance releases. New switch firmware
can be installed from the following host operating systems:
•UNIX
•Windows NT
•Windows 98
•Windows 95
Utility programs to facilitate loading firmware, sample Fabric Watch configurations, and MIB files
for switch management by SNMP can be accessed on the Brocade website through the following
steps:
1.Open a web browser at:
http://www.brocade.com
2.Click Support.
3.Click MIBs a nd RSH Utilities.
4.Click the link for the desired pro duct.
ivFabric Watch User’s Guide
Page 7
Introducing Brocade Fabric Watch
This chapter provides the following information:
•Overview on page 1-1
•Threshold Behavior Models on page 1-2
Overview
Chapter
1
Brocade Brocade Fabric Watch
elements, making it easy to quickly identify and escalate potential problems. It monitors each
element for out-of-boundary values or counters and provides notification when any exceed the
defined boundaries. The SAN manager can configure which elements, such as error, status, and
performance counters within a SilkWorm switch, are monitored.
Brocade Fabric W atch
switches and above with Brocades Fabric OS™ version v3.0. Bro cade Fabric Watch
accessed through:
TM
TM
is used by SAN managers to monitor key fabric and switch
is an optionally licensed product that runs on SilkWorm 2000 series
TM
can be
•Brocade Web Tools
•A telnet interface
•A SNMP-based enterprise manager
•By modifying and uploading the Brocade Fabric Watch
TM
Brocade Fabric W atch
monitors the following elements:
TM
configuration file to the switch.
•Fabric events (such as topology reconfigurations, zone changes)
•Switch environment (fans, power supplies, and temperature)
•Ports (state changes, errors, and performance)
•GBICs (for switches equipped with SMART GBICs.
Example: the Finiasr SMART GBIC FTR-8519-3).
With Brocade Fabric Watch
counters against a set of defined ranges. This and other information specific to each monitored
element is made available by Brocade Fabric Watch
modification. This set of information about each element is called a threshold, and the upper and
lower limits of the defined ranges are called boundaries.
TM
installed, each switch continuous ly mon itors err or and performance
TM
for viewing and, in some cases,
If conditions break out of acceptable rang es, an event is considered to have occurred, and one or
more of the following alarms (reporting mechanisms) are generated if configured for the relevant
threshold:
•SNMP trap
Fabric Watch User’s Guide1-1
Page 8
1
Introducing Brocade Fabric Watch
•Entry in the switch event log
•Locking of the port log to preserve the relevant informati on
•RAN (Remote Asynchronous Notification) Alert.
Note: The RAN alert is only available if the optional Fabric Access license has
been purchased and turned on.
•Email Alert
You can deploy Brocade Fabric Watch
profile from the following website:
http://www.brocade.com/support/mibs_Rsh
TM
Threshold Behavior Mode ls
The behavior models for th resholds are:
•Range
•Rising/falling
•Change monit or
as shipped, or download a customized configuration
Range Threshold
A range threshold tracks whether a fabric element is within a specified range. It includes a
minimum and maximum bound ary for the area, with buffer zones to prevent repeated events due to
oscillation of the value over a threshold boundary. If the value exceeds the low or high threshold
boundary, an event is generated. Fabric Watch can also generate events while the value is outside
the limits or when it re-enters the prescribed range.
1-2Fabric Watch User’s Guide
Page 9
Introducing Brocade Fabric Watch
Figure 1-1 shows a typical example of an error counter with a temperature range threshold.
1
Figure 1-1Example of Range Threshold: Temperature (Celsius)
Rising/Falling Threshold
A rising/falling thresh old tracks whether an element is on the desi red side of a boun dary. It includes
an upper and lower boundary, and the buffer zones are always zero. Events can be selected for
transitions between the boundaries. Rising/falling thresholds are typically used for rate based
counters. Rate based counters are all counters with a time base set to something other than “none”,
this includes the TX Perf and RX Perf counters.
Fabric Watch User’s Guide1-3
Page 10
1
Introducing Brocade Fabric Watch
Figure 1-2 shows a typical example of an error counter with a rising/falling threshol d.
Figure 1-2Example of Rising/Falling Threshold: Error Rate
Change Monitor Threshold
A change monitor threshold generates events whenever a counter value changes, regardless of the
type of change. This type of threshold is usually used to indicate state changes, such as zoning
changes.
Note:The change monitor counter value displayed is the cumulative count (historical count) and
not the delta value from the previous triggered value.
An illustration is not provided because change monitor thresholds do not include boundaries.
1-4Fabric Watch User’s Guide
Page 11
Installing Fabric Watch
This chapter provides the following information:
•Installing through telenet commands 2-1
•Installing through Brocade Web Tools
Each switch running Fabric Watch must have a Fabric Watch license installed on it. A license may
have been installed in the switch at the factory. If not, contact your switch supplier to obtain a
license key.
Fabric Watch requires a SilkWorm 2000 series switch or above running Brocade’s Fabr ic OS
3.0. A Fabric Watch license can be installed either using telnet commands or throu gh Brocade Web
TM
Tools
Installing through Telnet
To install Fabric Watch using telnet commands perform the following steps:
.
TM
2-2
Chapter
2
TM
1.Log onto the switch by telnet (refer to Brocade Fabric OS User’s Guide for details), using an
account that has administrative privileges.
2.To determine whether a Fabric Watch license is already installed on the switch, type
licenseShow on the telnet command line.
A list displays all the licenses currently installed on the switch.
Example:
admin> licenseShow
1A1AaAaaaAAAA1a:
Release v3.0
Web license
Zoning license
SES license
QuickLoop license
If the Fabric Watch license is not included in the list or is incorrect, continue with step [3].
3.Enter the following on the command line:
licenseAdd “key”
where “key” is the license key p rov i ded to you, surrounded by double quotes. The license key
is case sensitive and must be entered exactly as given.
4.Verify the license was added by entering the following on the command line:
licenseShow
If the license is not listed, repeat step [3].
Fabric Watch User’s Guide2-1
Page 12
2
Installing Fabric Watch
Installing throu gh Brocade Web Tools
To install Fabric Watch using Web Tools, perform the following steps:
1.Launch the web browser, enter the switch name or IP address in the Location/Address field
(for example:
Brocade Web Tools
2.Click the Admin button on the relevant switch panel.
The logon window displays.
3.Enter a logon name and password with administrative privileges and press Ent er.
The Administration View displays.
4.Select the License Admin tab, enter the license key in the License Key: field, and click Add License.
http://111.222.33.1), and press Enter.
TM
launches, displaying the Fabric View.
2-2Fabric Watch User’s Guide
Page 13
Using Fabric Watch
This chapter provides the following information:
• Overview on page 3-1
• User Interfaces on page 3-1
• Profiles on page 3-2
• Classes on page 3-3
• Threshold Naming Conventions on page 3-7
• Events on page 3-8
• Alarms on page 3-9
• Configuring Thresholds and Alarms on page 3-10
Overview
Chapter
3
Fabric Watch provides the following information about each out-of-boundary condition discovered:
•Name of threshold
•Current value of element counter
•Unit of measurement (for example, degrees Celsius, RPM, or unit of time)
•Time base for counter, used to compute rate of change (for example, events per minute)
•Historical information about the last alarmed event generated
User Interfaces
Use either Brocade Web ToolsTM, telnet interface, an SNMP-based enterprise manager, or the
configuration file to view and modify Fabric Watch settings.
Brocade Web Tools
Through Brocade Web ToolsTM, you can:
•View fabric and switch events through the fabric-wide Event View.
•View and modify threshold and alarm configurations through the Fabric Watch View.
•Upload and download the configuration file through the Config Admin tab in the Switch
Admin window.
Refer to the Brocade Web Tools User’s Guide for information on using Brocade Web Tool
TM
.
Fabric Watch User’s Guide3-1
Page 14
3
Using Fabric Watch
Telnet Interface
You can do the following using a telnet interface:
• Query fabric and switch events through the telnet command
• Query and modify threshold and alarm configurations. Use the the telnet command
fwConfigure to view the default and customized settings.
• Upload and download the configuration file through the commands
configDownload.
fwShow.
configUpload and
SNMP-Based Enterprise Managers
The Fabric Watch con figu ration information is stored as MIB variables, allowing you to do the
following:
• Query the MIB variable for individual fabric and switch elements.
• Query and modify threshold and alarm configurations.
• Receive generated SNMP traps when threshold conditions are met.
Profiles
Configuration File
You can view and modify the threshold and alarm configurations by uploading the configuration
file from the switch to the host, editing it in a text editor, then downloading the modified file back
to the switch. You can then ensure a uniform configuration throughout the fabric by distributing the
configuration file to all the switches in the fabric.
The configuration file can be uploaded and downloaded through either Brocade Web Tools
Config Admin tab in the Switch Admin window) or the telnet commands
configDownload. After downloading the file back to the switch, you must either reboot the
switch or use the telnet command
Each Fabric Watch profile co nt ains a uniq ue thresh old and alarm settin gs. Current profiles are
divided into:
fwConfigReload to reload the configuration file.
configUpload and
•debug
• workgroup errlog
•workgroup snmp
•enterprise errlog
•enterprise snmp
TM
(the
In addition, a Fabric W atch p rofile specific to your environment can be downloaded to the s witch to
customize the configuration.
3-2Fabric Watch User’s Guide
Page 15
Using Fabric Watch
Several customized profiles for Fabric Watch are available at the following website:
Each profile is for a different purpose, and contains only the configuration information that is
different from the default configuration.
The intention is to have distinct configuration setting for enterprise SAN and a workgroup SAN’s
are left for individual user’s.
The default information is always pres erved . Af ter load ing the p rof ile, yo u can s pecify wheth e r the
default or the profile information is used.
A profile may contain the following information:
3
•Threshold values:
- Unit string
- Tim e base
- Lower boundary
- Upper boundary
- Buffer size
•Alarm values
- Change
- Exceed
- Below
- Above
- In-between
- Behavior type
Classes
Fabric and switch elements are org anized into classes (also known as Agents), which are groupings
of closely related elements. There major classes are:
FabricMonitors key fabric resources, such as fabric reconfiguration,
zoning changes, new fabric logins
EnvironmentalMonitors switch environment functions, such as temperature,
power supply, fan status
PortMonitors port error and performance counters
E_PortMonitors E_Port error and performance counters
F/FL_Port (Optical)Monitors optical FL/F_Port error and performance counters
F/FL_Port (Copper)Monitors copper FL/F_Port error and performance counters
GBICMonitors operational values for FINISAR smart GBICs
ALPA
Note:The ALPA class is only available if the optional
Performance Monitoring license has been purchased and
turned on.
Fabric Watch User’s Guide3-3
Page 16
3
Using Fabric Watch
Monitors ALPA devices of Loop ports (currently CRC error of
ALPA device only).
End-to-End
1 - Monitors transmitting performance predefined by the
SID-DID pair. Configurable with either telnet commands or
Web Tools at run time.
2 - Monitors receiving performance predefined by the SID-DID
pair. Configurable with either telnet commands or Web Tools
at run time.
3 - Monitors CRC errors of prefedined S ID-DID pair.
Configurable with either telnet commands or Web Tools at run
time.
The configuration information of the End-to-End class is lost when
the switch is rebooted or fast booted.
Note:The End-to-End class is only available if the optional
Performance Monitoring license has been purchased and
turned on.
Note:The EE filter is preconfigured with either a telnet command
or Web Tools at run time.
Filter
Monitors the filter counter for a given port. Provides the following
types of information about the filter hit:
1 - Read command
2 - SCSI or IP traffic
3 - SCSI Read/Write
Note: The Filter class is only available if the optional
Performance Monitoring license has been purchased and
turned on.
Note:The type of filter is predefined by configuring it with either
a telnet command or Web Tools at run time.
In addition, each class is subdivided into areas, as listed in the following table.
PortLink failure countMonitors link failure count for each
E_PortLink failure countMonitors link failure of each E_port.
Zoning changesMonitors changes to currently
enabled zoning configuration.
Fabric to QuickLoop
changes
Fabric loginsMonitors number of host device
GBIC changeMonitors insertion/removal of GBIC.
FanMonitors operation of switch fans.
Power supplyMonitors status of each power supply.
Loss of synchronization
count
Loss of signal countMonitors port signal loss.
Primitive sequence
protocol error
Invalid transmission wordMonitors port invalid words.
Invalid CRC countMonitors port CRC errors.
Receive performanceMonitors port receive performance.
Transmit performanceMonitors port transmit performance.
State changesMonitors port state changes.
Loss of synchronization
count
Loss of signal countMonitors E_port signal loss.
Primitive sequence
• Abbreviation for the area name (alpha characters, title case). For example, “Temp” for the
T emp erature area.
• Index number for the number of the item within the series. Consists of three numbers, for
example: 000 for the first port, 001 for the next, etc. Index numbers for the Fabric, Port, E_Port,
F/FL_Port (Optical), F/ FL _Port (Copper), and GBIC classes begin with 000, and index numbers
for the Environment class begin with 001.
Example of a Threshold Name:
The threshold corresponding to the first thermometer in the switch is in the Environment class,
Temperature area, and is therefore named
An event is generated each time a boundary , as defin ed by the threshold, is cros sed. Bound aries are
not inclusive, so events are generated only when a boundary is exceeded, not when the monitored
value has only reached them. If the event has an assigned alarm, an alarm is also generated. The
alarm can be designated as an SNMP trap, an entry in the switch error log, locking of the port log,
or a combination of these options.
When an item such as an E_Port, F/FL_Port (optical), F/FL_Port (copper), FINISAR Smart GBIC,
fan, or power supply is removed, Fabric Watch may raise an event (such as a below event), then the
threshold is hidden and disabled. When an item is added, the threshold is displayed and enabled,
and Fabric Watch may raise an event.
envTemp001.
Event policies control the generation of events, and can be configured for either triggered events or
continuous events.
Triggered Events
A triggered event results in a single even t when a boundary is exceeded. The ev ent is not generated
again until the threshold value has returned within the boundaries and then once again exceeded
them. For example, if the switch temperature exceeds the upper boundary, a triggered event is
generated at the point the boundary is crossed, but is not repeated while the temperature remains
above the upper boundary.
The following events can be generated as triggered events:
StartedNo alarm is generated.
BelowCounter is below the lower boundary. Must be preceded by a start, above, or in-
between event.
AboveCounter is above the upper boundary. Must be preceded by a start, below, or in-
between event.
ExceededCounter is below the l ower bo undary or abo ve the up per boundar y. Accompanies
a below or above event.
ChangedCounter value has changed.
In-betweenCounter falls below the upper boundary minus buffer, or rises above the lower
boundary plus buff er. Must be preceded by an above or below event. If the bu ff er
is set to zero, this event is suppressed.
3-8Fabric Watch User’s Guide
Page 21
Alarms
Using Fabric Watch
3
Continuous Events
A continuous event results in an event at each time interval from when the boundary is initially
exceeded until the threshold value has returned within the boundaries. For example, if port
utilization is above the upper boundary, a new event is generated at each behavior interval until
utilization falls below the upper boundary. The following events can be generated as continuous
events:
StartedNo alarm is generated.
BelowCounter is below the lower boundary.
AboveCounter is above the upper boundary.
ExceededCounter is below the lower boundary or above the upper boundary.
Accompanies a below or above event.
ChangedCounter value has changed.
Each event can generate one or more alarms. Fabric Watch supports the following types of alarms:
•SNMP trap
•Switch event log entry
•Locking of the port log
•RAN
•Email
SNMP Trap
The following information is forwarded to an SNMP management station:
•Name of the element
•Class, area, and index of the threshold
•Ty pe of event generated
•Element value
•New state of element
Error Log Entry
The internal error log maintains a record of the event, up to a maximum of 64 entries. If configured
to do so, error log entries are forwarded to the SYSLOGD facility.
Locking the Port Log
Locking the Port Log freezes the switch port log to retain detail information about a problem. It is
typically used in conjunction with the error log entry.
Fabric Watch User’s Guide3-9
Page 22
3
Using Fabric Watch
RAN
RAJ sends alarm messages to the remote agent proxy server.
Note: The RAN alert is only available if the optional Fabric Access license has been purchased
and turned on.
Email Alert
The Email Alert sends alarms to the corresponding email addresses at run time. The configured
email address applies to all areas within a class. See fwMailCfg in the Fabric OS Reference Guide.
Note:This alert only works with firmware images that supports Web Tools.
Note:Whenever the switch is fastbooted or rebooted the Email alert is disabled. The Email Alert
must be reset after the switch is fastbooted or rebooted.
Configuring Thresholds and Alarms
The configuration of thresholds and alarms can b e divided into two categories: thr eshold values and
threshold area values.
Note:Timebase changes to thresholds in Env and GBIC agents are disabled. Non-factory
timebase settings in configuration files for both of these thresholds will not have any effect.
The factory timebase will be used.
Threshold Values
Threshold values apply to the specific threshold. They are not stored in the configuration file, and
return to the default values if the switch is rebooted. The following threshold values can be
modified:
StatusCan be enabled or disabled. Enabled by default.
Behavior typeAllows setting of the event policy to triggered or continuous. Set to
triggered by default.
Behavior intervalThe interval between the same type of alarm. Applies only to
continuous events. The default interval is 1 second.
3-10Fabric Watch User’s Guide
Page 23
Using Fabric Watch
3
Threshold Area Values
Threshold area values include boundaries an d alarms, and apply to all the thres holds within an area.
Changes are stored in the configuration file.
Boundaries
The following boundary information can be modified:
Unit stringRepresents unit value. Only the default unit string s are supported by
Fabric Watch.
Time baseTime period within which a specified event is measured. Can be
from one second to one day. Shorter time periods are more sensitive
to fluctuations and therefore will provide more detailed
information.
Note:The Time base boundary can only be modified through a
telnet command. The Time bas e bounder does not apply to
the:
- Transmit performance area
- Receive performance area
- Filter class
Low boundaryMinimum value. An event is generated if the element value falls
below this boundary.
High boundaryMaximum value. An event is generated if the element value rises
above this boundary.
Buffer sizeSize of buffer set up to decrease generation of in-between events
due to oscillation of the element value over a boundary.
Alarms
The following alarms can be added or deleted:
ERRLOGLogs errors to switch. If configured properly, sends message to
syslog on daemon.
SNMP-TRAPSends trap to SNMP agent.
PORT-LOG-LOCKFreezes port log to preserve log information generated at time of
event, for diagnostic purposes.
RANSends alarm messages to remote agent.
Note: The RAN alert is only available if the optional Fabric Access
license has been purchased and turned on.
Email AlertSends alarm messages to the corresponding mail address.
Fabric Watch User’s Guide3-11
Page 24
3
Using Fabric Watch
3-12Fabric Watch User’s Guide
Page 25
Example Error Message
This appendix provides an example Errlog message.
Overview
Figure A-1 shows a typical Fabric Watch Errlog message.
Appendix
A
Severity of Error
Message
3 = W arn ing
4= Informationa l
0x1003f2d0 (tThad): May 22 19:20:23
Error FW-BELOW, 3, fopportState003 (
. curren t value : 0
0x1003f2d0 (tThad): May 22 19:13:57
Error FW-ABOVE, 3, fopportCRCs007
current valu e : 5 Error(s)/ m i n u t e. (
0x1003f2d0 (tThad): May 22 17:31:33
Error FW-ABOVE, 3 , fopportSync003
value : 3 Error(s)/minute. (faulty)
0x1003f2d0 (tThad): May 22 17:31:33
Error FW-ABOVE, 3,fopportLink003
current value : 1 Error(s)/minu te
0x1003f2d0 (tThad): May 22 17:31:33
Error FW-CHANGED, 4,fabricFL000
value : 23 Login(s). (info)
0x1003f2d0 (tThad): May 22 17:17:33
Error FW-ABOVE, 3, alpaPerfCRC004
Change(s)/minute. (normal)
Threshold
Names
faulty
(
(faulty)
(
!!
Threshold
Labels
(
)
4
4) is above high boundary.
Index
Numbers
) is below low boundary
3
) is above high boundary.
7
) is above high boundary. current
3
) is above high boundary.
3
) value has changed. current
(faulty)
States
Figure A-1 Typical Fabric Watch Errlog Message
Fabric Watch User’s GuideA-1
Page 26
A
Example Error Message
Note:Threshold name = class name + area name + element index.
Invalid CRCs: CRCs
Link Failures: Link
Protocol Errors: ProtoErr
RX Performance: RXPerf
Loss of Signal: Signal
State Changes: State
Loss of Sync: Sync
TX Performance: TXPerf
Invalid Words: Words
E-port down: ED
Fabric login: FL
Fabric<->QL: FQ
Reconfigure: FR
GBIC change: GS
Segmentation: SC
Zoning change: ZC
Name server login: NL
Name server request: NR
Performance alpa monitor: alpaInvalid CRCs: PerfCRC
Performance end to end monitor: EEInvalid CRCs: PerfCRC
An encoding scheme that converts each 8-bit byte into 10 bits. Used to balance
ones and zeros in high-speed transports.
Address
Identifier
AL_PA
Alias Address
Identifier
Alias AL_PA
Alias Server
ANSI
API
Arbitrated
Loop
ASIC
ATM
AW_TOV
Bandwidth
BB_Credit
Beginning
Running
Disparity
A 24-bit or 8-bit value used to identify the source or destination of a frame.
Arbitrated Loop Physical Address; a unique 8-bit value assigned during loop
initialization to a port in an arbitrated loop.
An address identifier recognized by a port in addition to its standard identifier. An
alias address identifier may be shared by multiple ports.
An AL_PA value recognized by an L_Port in addition to the AL_PA assigned to
the port. See also AL_PA.
A fabric software facility that supports multicast group management.
American National Standards Institute; the governing body for fibre channel
standards in the U.S. A.
Application Programming Interface; defined protocol that allows applications to
interface with a set of services.
A shared 100 MBps fibre channel transport structured as a loop. Can support up
to 126 devices and one fabric attachment. See also Topology.
Application Specific Integrated Circuit.
Asynchronous Transfer Mode; a transport used for transmitting data over LANs
or WANs that transmit fixed-length units of data. Provides any-to-any
connectivity, and allows nodes to transmit simultaneously.
Arbitration Wait Time-out Value; the minimum time an arbitrating L_Port waits
for a response before beginning loop initialization.
The total transmission capacity of a cable, link, or system. Usually measured in
bps (bits per second). May also refer to the range of transmission frequencies
available to a network. See also Throughput.
Buffer-to-buffer credit; the number of frames that can be transmitted to a directly
connected recipient or within an arbitrated loop. Determined by the number of
receive buffers available. See also Buffer-to-buffer Flow Control, EE_Credit.
The disparity at the transmitter or receiver when the special character associated
with an ordered set is encoded or decoded. See also Disparity.
BER
Bit Error Rate; the rate at which bits are expected to be received in error.
Expressed as the ratio of error bits to total bits transmitted. See also Error.
Block
As applies to fibre channel, upper-level application data that is transferred in a
single sequence.
Fabric Watch User’s GuideGlossary-1
Page 30
Broadcast
Brocade®
Distributed
Fabrics
The transmission of data from a single source to all devices in the fabric,
regardless of zoning. See also Multicast, Unicast.
The combined manual for the software products Brocade Extended Fabrics and
Brocade Remote Switch (“Distributed Fabrics” is not a software product). See
also Brocade Extended Fabric s, Brocade Remote Switch.
Brocade
Extended
Fabrics
TM
Brocade Fabric
Access
TM
Brocade Fabric
TM
Assist
Brocade Fabric
Manager
TM
Brocade Fabric
TM
Watch
Brocade ISL
Trunking
TM
Brocade
Performance
Monitoring
TM
A Brocade product that runs on Fabric OS and allows creation of a fibre channel
fabric interconnected over distances of up to 100 kilometers. See also Brocade Distributed Fabrics.
A Brocade product that consists of a set of APIs that allow third party applications
to interface with Fabric OS.
A Brocade feature that enables private and public hosts to access public targets
anywhere on the fabric, provided they are in the same Fabric Assist zone. This
feature is available only when both Brocade QuickLoop and Brocade Zoning are
installed on the switch.
A Brocade product that works in conjunction with Web Tools to provide a
graphical user interface for managing s witch group s (such as th e SilkWorm 6400)
as a single unit, instead of as separate switches. Fabric Manager is installed on
and run from a computer workstation.
A Brocade product that runs on Fabric OS and allows monitoring and
configuration of fabric and switch elements.
A Brocade feature that enables distribution of traffic over the combined
bandwidth of up to four ISLs (between adjacent switches), while preserving inorder delivery. A set of trunked ISLs is called a trunking group; each port
employed in a trunking group is called a trunking port. See also Master Port .
A Brocade product that provides error and performance information to the
administrator and end user for use in storage management.
Brocade
QuickLoop
A Brocade product that makes it possible to allow private devices within loops to
TM
communicate with public and private devices across the fabric through the
creation of a larger loop.
May also refer to the arbitrated loop created using this software. A QuickLoop
can contain a number of devices or looplets; all devices in the same QuickLoop
share a single AL_PA space.
Brocade
Remote
Switch
TM
A Brocade product tha t run s on Fabric OS and enables two S ilkWorm 2000 fabric
switches to be connected over an ATM (asynchronous transfer mo de) connecti on.
This requires a compatible fibre channel to ATM gateway , and can have a dis tance
of up to 10 kilometers between each switch and the respective ATM gateway. See
also Brocade Distributed Fabrics.
Brocade SES
TM
A Brocade product that runs on Fabric OS and allows monitoring, configuring,
and maintenance of the Brocade SilkWorm Switch family using SCSI-3
Enclosure Services.
Brocade Web
TM
Tools
A Brocade product that runs on Fabric OS and prov ides a graphical interface to
allow monitoring and manag ement of indi vi du al sw it ches or en tir e fabri cs fro m a
standard workstation.
Glossary-2Fabric Watch User’s Guide
Page 31
Brocade
Zoning
Buffer-to-buffer
Flow Control
TM
A Brocade product that runs on Fabric OS and allows partitioning of the fabric
into logical groupings of devices. Devices in a zone can only access and be
accessed by devices in the same zone. See also Zone.
Management of the frame transmission rate in either a point-to-point topology or
in an arbitrated loop. See also BB_Credit.
Cascade
T wo or mo re interconnected fi bre channel switches. SilkWorm 2000 switches can
be cascaded up to 239 switches, with a recommended maximum of seven
interswitch links (no path longer than eight switches). See also Fabric, ISL.
Chassis
Circuit
The metal frame in which the switch and switch components are mounted.
An established communication path between two ports. Consists of two virt ual
circuits capable of transmitting in opposite directions. See also Link.
Class 1
Service that provides a dedicated connection between two po rts (also called
connection-oriented service), with notification of delivery or nondelivery.
Class 2
Service that provides multiplex and connectionless frame switching service
between two ports, with notification of delivery or nondelivery.
Class 3
Service that provides a connectionless frame switching service between two ports,
without notification of delivery or nondelivery of data. Can also be used to
provide a multicast connection between the originator and recipients, with
notification of delivery or nondelivery.
Class F
Connectionless service for control traffic between switches, with notification of
delivery or nondelivery of data between the E_Ports.
Class of
Service
Comma
A specified set of delivery characteristics and attributes for frame delivery.
A unique pattern (eith er 1 100000 or 001 1111) used i n 8B/10B encodi ng to specify
character alignment within a data stream. See also K28.5.
Community
(SNMP)
A relationship between a group of SNMP managers and an SNMP ag ent, in which
authentication, access control, and proxy characteristics are defined. See also
SNMP.
CRC
Cyclic Redundancy Check; a check for transmission errors included in every data
frame.
Credit
As applies to fibre channel, the number of receive buffers available for
transmission of frames between ports. See also BB_Credit, EE_Credit.
Cut-throughA switching technique that allows the route for a frame to be selected as soon as
the destination address is received. See also Route.
Data Word
Type of transmission word that occurs within frames. The frame header, data
field, and CRC all consist of data words. See also Frame, Ordered set, Transmission Word.
Defined Zone
Configuration
The set of all zone objects defined in the fabric. May include multiple zone
configurations. See also Enabled Configuration, Zone Configuration.
Disparity
The relationship of ones and zeros in an encoded character. “Neutral disparity”
means an equal number of each, “positive disparity” means a majority of ones,
and “negative disparity” means a majority of zeros.
DLS
Dynamic Load Sharing; dynamic distribution of traffic over available paths.
Allows for recomputing of routes when an Fx_Port or E_Port changes status.
Fabric Watch User’s GuideGlossary-3
Page 32
Domain ID
E_D_TOV
E_Port
EE_Credit
EIA Rack
Enabled Zone
Configuration
As applies to SilkWorm switches, a unique number between 1 and 239 that
identifies the switch to the fabric and is used in routing frames. Usually
automatically assigned by the switch, but can be manually assigned.
Error Detect Time-out Value; the minimum amount of time a target waits for a
sequence to complete before initiating recovery. Can also be defined as the
maximum time allowed for a round-trip transmission before an error condition is
declared. See also R_A_TOV, RR_TOV.
Expansion Port; a type of switch port that can be connected to an E_Port on
another switch to create an ISL. See also ISL.
End-to-end Credit; the number of receive buffers allocated by a recipient port to
an originating port. Used by Class 1 and 2 services to manage the exchange of
frames across the fabric between source and destination. See also End-to-end Flow Control, BB_Credit.
A storage rack that meets the standards set by the Electronics Industry
Association.
The currently enabled configuration of zones. Only one configuration can be
enabled at a time. See also Defined Configuration, Zone Configuration.
End-to-end
Flow Control
Error
Exchange
F_Port
Fabric
Fabric Name
Fabric OS
FC-AL-3
FC-FLA
FCP
FC-PH-1, 2, 3
TM
Governs flow of class 1 and 2 frames between N_Ports. See also EE_Credit.
As applies to fibre channel, a missing or corrupted frame, time-out, loss of
synchronization, or loss of signal (link errors). See also Loop Failure.
The highest level fibre channel mechanism used for communication between
N_Ports. Composed of one or more related sequences, and can work in either one
or both directions.
Fabric Port; a port that is able to transmit under fabric protoco l and interface o ver
links. Can be used to connect an N_Port to a switch. See also FL_Port, Fx_Port.
A fibre channel net work con t aini ng t wo o r more switches in addition to hosts and
devices. May also be referred to as a switched fabric. See also Topology, SAN, Cascade.
The unique identifier assigned to a fabric and communicated during login and
port discovery.
The proprietary operating system on Brocade switches.
The Fibre Channel Arbitrated Loop standard defined by ANSI. Defined on top of
the FC-PH standards.
The Fibre Channel Fabric Loop At tach standard defined by ANSI.
Fibre Channel Protocol; mapping of protocols onto the fibre channel standard
protocols. For example, SCSI FCP maps SCSI-3 onto fibre channel.
The Fibre Channel Physical and Signalling Interface standards defined by ANSI.
FC-PI
FC-PLDA
The Fibre Channel Physical Interface standard defined by ANSI.
The Fibre Channel Private Loop Direct Attach standard defined by ANSI. Applies
to the operation of peripheral devices on a private loop.
Glossary-4Fabric Watch User’s Guide
Page 33
FC-SW-2
Fibre Channel
Transport
The second generation of the Fibre Channel Switch Fabric standard defined by
ANSI. Specifies tools and algorithms for the interconnection and initialization of
fibre channel switches in order to create a multi-switch fibre channel fabric.
A protocol service that supports communication between fibre channel service
providers. See also FSP.
Fill Word
Firmware
FL_Port
FLOGI
Frame
FRU
FS
FSP
FSPF
Full-duplex
Fx_Port
G_Port
GBIC
Gbps
An IDLE or ARB ordered set that is transmitted during breaks between data
frames to keep the fibre channel link active.
The basic operating system provided with the hardware.
Fabric Loop Port; a port that is able to transmit under fabric protocol and also has
arbitrated loop capabilities. Can be used to connect an NL_Port to a switch. See
also F_Port, Fx_Port.
Fabric Login; the process by which an N_Port determines whether a fabric is
present, and if so, exchanges service parameters with it. See also PLOGI.
The fibre channel structure used to transmit data between ports. Consists of a
start-of-frame delimiter, header, any optional headers, the data payload, a cyclic
redundancy check (CRC), and an end-of-frame delimiter. There are two types of
frames: Link control frames (transmission acknowledgements, etc.) and data
frames.
Field-replaceable Unit; a component that can be replaced on site.
Fibre Channel Service; a service that is defined by fibre channel standards and
exists at a well-known address. For example, the Simple Name Server is a fibre
channel service. See also FSP.
Fibre Channel Service Protocol; the common protocol for all fabric services,
transparent to the fabric type or topology. See also FS.
Fabric Shortest Path First; Brocade’s routing protocol for fibre channel switches.
A mode of communication that allows the same port to simultaneously transmit
and receive frames. See also Half-duplex.
A fabric port that can operate as either an F_Port or FL_Port. See also F_Port,
FL_Port.
Generic Port; a port that can operate as either an E_Port or F_Port. A port is
defined as a G_Port when it is not yet connected or has not yet assumed a specific
function in the fabric.
Gigabit Interface Converter; a removable serial transceiver module that allows
gigabaud physical-level transport for fibre channel and gigabit ethernet.
Gigabits per second (1,062,500,000 bits/second).
GBps
Half-duplex
GigaBytes per second (1,062,500,000 bytes/second).
A mode of communication that allows a port to either transmit or receive frames
at any time, but not simultaneously (with the exception of link control frames,
which can be transmitted at any time). See also Full-duplex.
Hard Address
HBA
The AL_PA that an NL_Port attempts to acquir e duri ng lo op ini tialization.
Host Bus Adapter; the interface card between a server or workstation bus and the
fibre channel network.
Fabric Watch User’s GuideGlossary-5
Page 34
Hub
Idle
Initiator
Integrated
Fabric
A fibre channel wiring concentrator that collapses a loo p to polo gy into a ph ys ical
star topology. Nodes are automatically added to the loop when active and
removed when inactive.
Continuous transmission of an ordered set over a fibre channel link when no data
is being transmitted, to keep the link active and maintain bit, byte, and word
synchronization.
A server or workstation on a fibre channel network that initiates communications
with storage devices. See also Target.
The fabric created by a SilkWorm 6400, consisting of six SilkWorm 2250
switches cabled together and configured to handle traffic as a seamless group.
IOD
ISL
Isolated
E_Port
IU
JBOD
K28.5
L_Port
Latency
Link
Link Services
In-order Delivery; a parameter that, when set, guarantees that frames are either
delivered in order or dropped.
Interswitch Link; a fibre channel link f rom the E_P ort of one swi tch to the E_Por t
of another. See also E_Port, Cascade.
An E_Port that is online but not operational due to overlapping domain IDs or
nonidentical parameters (such as E_D_TOVs). See also E_Port.
Information Unit; a set of information as defined by either upper-level process
protocol definition or upper-level protocol ma pping.
Just a Bunch Of Disks; indicates a number of disks connected in a single chassis
to one or more controllers. See also RAID.
A special 10-bit character used to indicate the beginning of a transmission word
that performs fibre chann el contr ol and s ignaling functions . The fi rst se ven bits of
the character are the comma pattern. See also Comma.
Loop Port; a node port (NL_Port) or fabri c port (FL_Por t) that has arbitrat ed loop
capabilities. An L_Port can be in one of two modes:
•Fabric mode Connected to a port that is not loop capable, and using fabric
protocol.
•Loop mode In an arbitrated loop and using loop protocol. An L_Port in loop
mode can also be in participat ing mode or non-participating mode.
See also Non-participating Mode, Participating Mode.
The period of time required to transmit a frame, from the time it is sent until it
arrives.
As applies to fibre channel, a physical connection between two ports, consisting
of both transmit and receive fibres. See also Circuit.
A protocol for link-related actions.
LIP
Loop Initialization Primitive; the signal used to begin initialization in a loop.
Indicates either loop failu re or resetting of a node.
LM_TOV
Loop Master Time-out Value; the minimum time that the loop master waits for a
loop initialization sequence to return.
Loop Failure
Loss of sign al within a loop for any period of time, or loss of synchronization for
longer than the time-out value.
Loop
Initialization
Glossary-6Fabric Watch User’s Guide
The logical procedure used by an L_Port to discover its en vironment. Can be used
to assign AL_PA addresses, detect loop failure, or reset a node.
Page 35
Loop_ID
Looplet
A hex value representing one of the 127 possible AL_PA values in an arbitrated
loop.
A set of devices connected in a loop to a port that is a member of another loop.
LPSM
LWL
Master Port
MIB
Multicast
Multimode
N_Port
Name Server
NL_Port
Node
Loop Port State Machine; the logical entity that performs arbitrated loop
protocols and defines the behavior of L_Ports when they require access to an
arbitrated loop.
Long Wavelength; a type of fiber optic cabling that is based on 1300mm lasers
and supports link speeds of 1.0625 Gbps. May also refer to the type of GBIC or
SFP. See also SWL.
As relates to trunking, the port that determines the routing paths for all traffic
flowing through the trunking group. One of the ports in the first ISL in the
trunking group is designated as the master port for that group. See also Brocade ISL Trunking.
Management Information Base; an SNMP structure to help with device
management, providing configuration and device information.
The transmission of data from a single source to multiple specified N_Ports (as
opposed to all the ports on the network). See also Broadcast, Unicast.
A fiber optic cabling specification that allows up to 500 meters between devices.
Node Port; a port on a node that can connect to a fibre channel port or to another
N_Port in a point-to-point connection. See also NL_Port, Nx_Port.
Frequently used to indicate Simple Name Server. See also SNS.
Node Loop Port; a node port that has arbitrated loop capabilities. Used to connect
an equipment port to the fabric in a loop configuration through an FL_Port. See
also N_Port, Nx_Port.
A fibre channel device that contains an N_Port or NL_Port.
Node Name
Non-
participating
Mode
The unique identifier for a node, communicated during l ogin and port discovery.
A mode in which an L_Port in a loop is inactive and cannot arbitrate or send
frames, but can retransmit any received transmissions. This mode is entered if
there are more than 127 devices in a loop and an AL_PA cannot be acquired. See
also L_Port, ParticipatingMode.
Nx_Port
Ordered Set
A node port that can operate as either an N_Port or NL_Port.
A transmission word that uses 8B/10B mapping and begins with the K28.5
character. Ordered sets occur outside of frames, and include the following items:
•Frame delimiters Mark frame boundaries and describe frame contents.
•Primitive signals Indicate events.
•Primitive sequences Indicate or initiate port states.
Ordered sets are used to differentiate fibre channel control information from data
frames and to manage the transport of frames.
Packet
Participating
Mode
Fabric Watch User’s GuideGlossary-7
A set of information transmitted across a network. See also Frame.
A mode in which an L_Port in a loop has a valid AL_PA and can arbitrate, send
frames, and retransmit received transmissions. See also L_Port, Non-participati ng Mo de .
Page 36
Path Selection
Phantom
Address
The selection of a transmission path through the fabric. B r ocade switches use the
FSPF protocol.
An AL_PA value that is assigned to an device that is not physically in the loop.
Also known as phantom AL _PA.
Phantom
Device
PLOGI
Point-to-point
Port_Name
POST
Private
NL_Port
Private Device
Private Loop
Protocol
Public
NL_Port
Public Device
Public Loop
R_A_TOV
RAID
Request Rate
A device that is not physically in an arbitrated loop but is logically included
through the use of a phantom address.
Port Login; the port-to-port login process by which initiators establish sessions
with targets. See also FLOGI.
A fibre channel topology that employs direct links between each pair of
communicating entities. See also Topology.
The unique identifier assigned to a fibre channel port. Communicated during
login and port discovery.
Power On Self-Test; a series of tests run by a switch after it is turned on.
An NL_Port that communicates only with other private NL_Ports in the same
loop and does not log into the fabric.
A device that supports arbitrated loop protocol and can interpret 8-bit addresses,
but cannot log into the fabric.
An arbitrated loop that does not include a participating FL_Port.
A defined method and a set of standards for communication.
An NL_Port that logs into the fabric, can function within either a public or a
private loop, and can communicate with either private or public NL_Ports.
A device that supports arbitrated loop protocol, can interpret 8-bit addresses, and
can log into the fabric.
An arbitrated loop that includes a participating FL_Port, and may contain both
public and private NL_Ports.
Resource Allocation Time-out Value; the maximum time a frame can be delayed
in the fabric and still be delivered. See also E_D_TOV, RR_TOV.
Redundant Array of Independent Disks; a collection of disk drives that appear as
a single volume to the server and are fault tolerant through mirroring or parity
checking. See also JBOD.
The rate at which requests arrive at a servicing entity. See also Service Rate.
Route
As applies to a fabric, the communication path between two switches. May also
apply to the specific path taken by an individual frame, from source to
destination. See also FSPF.
Routing
RR_TOV
The assignment of frames to specific switch ports, according to fr ame destination .
Resource Recovery Time-out Value; the minimum time a target device in a loop
waits after a LIP before logging out a SCSI initiator. See also E_D_TOV, R_A_TOV .
RSCN
Registered State Change Notification; a switch function that allows notification of
fabric changes to be sent from the switch to specified nodes.
SAN
Storage Area Network; a network of systems and storage devices that
communicate using fibre channel protocols. See also Fabric.
Glossary-8Fabric Watch User’s Guide
Page 37
Sequence
A group of related frames transmitted in the same direction between two N_Ports.
Service Rate
SFP
SI
SilkWorm
Single Mode
SNMP
SNS
Switch
Switch Name
Switch Port
The rate at which an entity can service requests. See also Request Rate.
Small Form Factor Pluggable; optical transceiver used to convert signals b etween
optical fiber cables and switches.
Sequence Initiative.
The brand name for Brocade’s family of switches
The fiber optic cabling standard that corresponds to distances of up to 10 km
between devices.
Simple Network Management Protocol. An internet management protocol that
uses either IP for network-level functions and UDP for transport-level functions,
or TCP/IP for both. Can be made available over other protocols, such as UDP/IP,
because it does not rely on the underlying communication protocols. See also
Community (SNMP).
Simple Name Server; a switch service that stores names, addresses, and attributes
for up to 15 minutes, an d p rov id es them as required to other devices in the fabric.
SNS is defined by fibre channel standards and exists at a well-known address.
May also be referred to as directory service. See also FS.
Hardware that routes frames according to f ibre channel protocol and is controlled
by software.
The arbitrary name assigned to a switch.
A port on a switch. Switch ports can be E_Ports, F_Ports, or FL_Ports .
SWL
Target
Tenancy
Throughput
Topology
Translative
Mode
Transmission
Character
Transmission
Word
Trap (SNMP)
Short W avelength; a type of fiber optic cabling that is based on 850mm lasers and
supports 1.0625 Gbps li nk s peeds. May also r efer to the type of G BIC or SFP. See
also LWL.
A storage device on a fibre channel network. See also Initiator.
The time from when a port wins arbitration in a loop until the same port returns to
the monitoring state. Also referred to as loop tenancy.
The rate of data flow achieved within a cable, link, or system. Usually measured
in bps (bits per second). See also Bandwidth.
As applies to fibre channel, the configur ation of the fibre channel network and the
resulting communication paths allowed. There are three possible topologies:
•Point to point - A direct link between two communication ports.
•Switched fabric - Multiple N_Ports linked to a switch by F_Ports.
•Arbitrated loop - Multiple NL_Ports connected in a loop.
A mode in which private devices can communicate with public devi ces across the
fabric.
A 10-bit character encoded according to the rules of the 8B/10B algorithm.
A group of four transmission characters.
The message sent by an SNMP agent to inform the SNMP management station of
a critical error. See also SNMP.
Fabric Watch User’s GuideGlossary-9
Page 38
Tunneling
U_Port
UDP
ULP
ULP_TOV
Unicast
Well-known
Address
A technique for enabling two networks to communicate when the source and
destination hosts are both on the same type of network, but are connected by a
different type of network.
Universal Port; a switch port that can operate as a G_Port, E_Port, F_Port, or
FL_Port. A port is defined as a U_Port when it is not connected or has not yet
assumed a specific function in the fabric.
User Datagram Protocol; a protocol that runs on top of IP and provides port
multiplexing for upper-level protocols.
Upper-level Protocol; the protocol that runs on top of fibre channel. Typical
upper-level protocols are SCSI, IP, HIPPI, and IPI.
Upper-level Time-out Value; the minimum time that a SCSI ULP process waits
for SCSI status before initiating ULP recovery.
The transmission of data from a single source to a single destination. See also
Broadcast, Multicast.
As pertaining to fibre channel, a logical addr ess defined by the fibre channel
standards as assigned to a specific function, and stored on the switch.
Workstation
WWN
Zone
Zone
Configuration
A computer used to access and manage the fabric. May also be referred to as a
management station or host.
Worldwide Name; an identifier that is unique worldwide. Each entity in a fabric
has a separate WWN.
A set of devices and hosts attached to the same fabric and configured as being in
the same zone. See also Brocade Zo ni ng. Devices an d hos ts within th e s ame zo ne
have access permission to others in the zone, but are not visible to any outside the
zone.
A specified set of zones. Enabling a configuration enables all zones in that
configuration. See also Defined Configuration, Enabled Configuration.