An Ethernet based alarm system
ArticleCategory: [Choose a category, do not translate
this]
Hardware
AuthorImage:[Here we need a little image from you]
TranslationInfo:[Author + translation history. mailto: or
http://homepage]
original in en Guido Socher
AboutTheAuthor:[A small biography about the author]
Guido likes Linux because it is a really good system to
develop your own hardware.
Abstract:[Here you write a little summary]
Many alarm system for buildings use dedicated wiring. They are
often just alarming and monitoring one single building.
Using Ethernet and IP we can re-use the existing network in a
building (local LAN) for the alarm system. This has
especially advantages when a whole campus with several buildings
or units in different locations needs to be monitored. For larger distances
the public internet can be used as a transport media.
The idea was brought up by a customer who wanted to build an
earthquake monitoring system which reports alarms from "P-wave"
sensors (seismic wave sensors) several hundred kilometers away to a central server.
From there actions can then be taken to e.g park elevators in
a save position and shut them off before the wave hits the building.
ArticleIllustration:[This is the title picture for your
article]
ArticleBody:[The article body]
How it works
Alarms are reported by little tuxgraphics ethernet boards to
a central sever. The boards do alarm on contact open or contact
close dependent on the needs. Each of the boards gets a name
that it reports together with the alarm message. This
will help to identify the cause/location. The name can e.g be
building "section-42". The central sever is e.g a linux
PC. All it needs is a network card and the ability to run perl code.
The actual server code is a small perl program. It was done
this way to make it easy to adapt the code.
Layout of the alarm system with ethernet boards connected via
different networks.
The central alarm server maintains not only a log of all the alarms but
has the possibility to execute other commands.
In the simplest possible case it would e.g update a web page with the
alarm information and send an email (e.g using the unix Mail command).
It can as well use other tuxgraphics ethernet boards with the "ethernet remote
switch application" software loaded to switch something on or off (e.g switch on
a siren).
Different alarm servers
In the download section you find as part of the avr ethernet board
code in the sub-directory "servers" three different servers.
- alarmsrv-basic -- a very simple server, just a few lines of code. Easy to
understand.
- alarmsrv-multialarm -- a server that triggers only when there are multiple
events from a number of servers (important for a reliable earthquake response).
- alarmsrv-multi-hb -- like the alarmsrv-multialarm but has a heartbeat
supervision of the ethernet boards.
Let's talk
The communication protocol for the alarm system is UDP. The main reason
for using UDP was speed. As I mentioned this was build as a earthquake
response system. The seismic waves travel roughly at the speed of sound.
There is no time for the 3-way handshake which is required to setup a TCP connection and then send
the actual data. You want a system that responds in a few milliseconds.
The internet's routers and server may as well fall apart just a few milliseconds
after the seismic wave hits the area where the sensor is.
By using UDP packets we can achieve very low latency.
Security within the
alarm system is maintained by using a shared secret (aka password) between
ethernet boards out in the field and the central server. Only messages
that contain a valid shared secret are processed by the central server.
This should provide a sufficient level of security for most applications.
The central alarm server has to be reachable via a static IP address.
DNS look-up would be way too time consuming. The server does not have
to physically sit in at an ISP. It could be connected even via DSL as
long as your ISP assigns you a static IP. Many regional ISPs do that
for a small extra fee (e.g 5 dollar per month).
Other applications
This alarm system has a few properties which make it usable in many
areas not only as a home alarm system or earthquake response system.
It is based on small robust and independent tuxgraphics ethernet boards
sitting somewhere. All they need is power and some kind of network
connectivity to a nearby IP network or DSL router. These boards
can then report events at the speed that IP packets travel through
the wire.
It's a simple, robust and flexible system made to be tailored according to your
ideas.
Applications for this could be not only alarm and
monitoring systems. You can use it e.g in a production system
to report statistics and environmental parameters.
User interface
The advantage of using an IP network for the connectivity of the alarm
sensors is that you can easily get a remote management interface.
We run on the ethernet board in addition to the alarm reporting software
also a web server. This gives us an easy to use management interface
for the remote units.
Remote management and configuration.
Checking the alarm system manually.
References/Download