| |
|
| |
Simple
Redundant Pathways |
| |
The simplest
form of redundancy is just having two separate connections between
managed switches. RSTP will assign the connection with the lowest
"path cost" (shortest, fastest and/or least expensive
route) as the primary pathway and the other will be "blocked"
but kept ready as the backup pathway. Typically, you should
route these two connections through different conduits to make
sure they aren't both cut if a conduit is damaged.

|
| |
|
| |
Single
Ring or Loop |
|
| |
The
most common redundant network topology with managed switches is
to form a ring or loop. Typically, the managed switches are connected
in series and the last switch is then connected back to the first.
In this scheme the RSTP will "block" one of the connections
between the switches and maintain it as a backup pathway. If a
switch goes down or a connection is lost the managed switch will
automatically and quickly re-configure to use the backup pathway(s).

|
| |
|
| |
Ring
Coupling |
|
| |
RSTP
allows for multiple loops to be formed and interconnected. A common
application for this is when you need to connect to managed switches
in a remote location of your site. This gives you both local loop
redundancy and redundant connections between your loops.

|
| |
|
| |
Dual
Rings or Loops |
|
| |
One
of the more advanced levels of network redundancy is to have two
independent Ethernet networks where each is connected in a loop.
This topology is ideal when you have advanced devices with dual
Ethernet ports such as the SixTRAK
IPm DCS Controller or EtherTRAK
I/O Gateway. Use this scheme in combination with redundant
controllers, redundant
I/O and redundant
supervisory HMI/SCADA computers to get the highest level of
system redundancy.

|
| |
|
| |
Mesh
Topology |
|
| |
In
this scheme every "designated" (non-root) switch has
a direct connection back to the "root" switch. In addition,
there are connections between the "designated" switches
creating redundant loops. The advantage to this topology over
just a simple loop is that it affords the maximum bandwidth because
of the direct connection to the "root". The loop connections
are only used as backups. This scheme is tolerant of multiple
switch or connection failures.

|
| |
|
| |
RSTP
Setup Guidelines |
|
| |
The
Rapid Spanning Tree Protocol may seem complicated but is actually
simple to setup and use. In most cases, where all you want is
a simple ring of switches, all you have to do is enable RSTP and
the rest is plug-and-play. You can leave all the RSTP parameters
at their factory defaults and the switch will automatically determine
the best primary and backup pathways. For the best RSTP performance
here are some basic guidelines to follow:
- Redundancy
protocol - Choose the Rapid Spanning Tree Protocol
(RSTP) for most applications because it offers the fastest recovery
times and is fully compatible with the original Spanning Tree
Protocol (STP). The SIXNET managed switch with firmware v2.6
or newer is based on the latest IEEE 802.1D-2004 standard for
RSTP and STP.
- Typical
recovery time - The SIXNET Managed Switches (firmware
v3.1 or higher), with RSTP enabled, typically provide less than
50 mS per hop recovery times when there is a link-loss failure
anywhere in the network. A typical ring network of ten managed
switches will provide a recovery time of <500 mS.
- Number
of switches in a ring - Based on the IEEE 802.1D-2004
standard the maximum number of "hops" in a ring topology
is 40. A hop is defined as a link between two switches. This
allows for up to 40 switches in a ring but better performance
is achieved with less switches.
- Bridge
priority
- This along with the MAC address determines which managed switch
is considered the "root bridge". For simple ring networks
(which are most common) you can leave all the switches with
the default bridge priority and they will automatically decide
which should be the "root" based on the MAC address.
You only need to change this value in advanced topologies where
you want to force a switch to be the root bridge.
- Port
priority - If a switch has redundant direct connections
to the root bridge then the port priority determines which will
be the primary and backup connections. In most cases you should
use the default setting (which is the same for all ports) and
let the switch automatically determine which port to make active
(port with lowest port number) and which to block.
- Maximum
age, hello time, forward delay & transmission limit
- These parameters define various aspects of how the RSTP operates.
It is recommended that you leave these parameters at the default
values for all the managed switches in your network.
- Path
cost - This parameter helps the switch decide which
port offers the best path to the root bridge. It is recommended
that you leave it at the default setting of Auto which will
let the switch automatically make the determination.
- Point-to-point
MAC - This parameter tells the switch that a port has
a direct (one-one-one) connection to another switch which allows
it to start forwarding traffic immediately. It is recommended
that you leave it at the default setting of Auto which will
let the switch automatically make the determination.
- Edge
port - If you know a port is directly connected to
an end device (that doesn't support RSTP) then set it as an
edge port to ensure maximum performance. This will tell the
switch to immediately start forwarding traffic on the port and
not bother trying to establish a RSTP connection.
|