Echelon ISI - Manual (Page 12)

ISI Programmer’s Guide 10
increases the network bandwidth consumed for administrative ISI messages,
possibly preventing regular network operation due to an increased collision rate.
Networks should be designed with some headroom. A reasonable limit is 80%.
ISI-S networks that reach 26 devices should be considered for an upgrade to ISI-
DA (which might be as simple as adding a DAS), and ISI-DA networks exceeding
160 devices are prime candidates for an upgrade to a managed LNS network.
Channel Types and Limits
The supported channel types for the ISI protocol are PL-20 power line and
TP/FT-10 free topology twisted pair.
The maximum channel limit for a device using the IsiCompactManual or
IsiCompactAuto library is one channel. For devices using any of the other
libraries, the limit is two channels. ISI-DAS devices always support two
channels.
The IsiCompactManual and IsiCompactAuto libraries are provided for feature-
limited implementations of ISI devices with minimum memory footprint. The
functionality of the ISI libraries is described in
Optimizing the Footprint of ISI
Applications .
Supported Topology and Routers
ISI-S networks are limited to a single channel that is segmented with physical
layer repeaters according to the standard channel properties—none for PL-20
channels, or multiple for TP/FT-10 channels provided there is never more than
one physical layer repeater between any two points of communication. In other
words, you can have one N-way repeater, much in the way of an N-port Ethernet
hub. A physical layer repeater is similar to a hub (signal booster without
filtering logic).
ISI-DA networks can have one or two channels. ISI-DA networks with two
channels must include a router configured as a repeater. Each channel must
meet the same requirements as a channel for ISI-S without a DAS described
above. The router must be preconfigured to be compatible with ISI networks, or
otherwise capable of joining an ISI network.
If a domain address server is used in a two-channel network with a PL-20 and
TP/FT-10 channel, it should be located on the PL-20 channel. One of the
functions of the domain address server is to determine the slowest channel of the
network that it is located on. If a domain address server is located on the PL-20
channel, it will start-up with knowledge of the slowest channel. If it is located on
the TP/FT-10 channel, it will have to learn of the existence of the PL-20 channel
by discovering one of the PL-20 devices. This may take some time. Conversely, if
the domain address server is located on the TP/FT-10 channel and all PL-20
devices are removed from the network, the domain address server should be reset
to relearn the network topology.