Sample 2 - Who-Is/I-Am Counter

This sample application builds on the first sample by overriding the default processing for Who-Is and I-Am requests, counting them, then continuing on with the regular processing.

The description of this sample will be about the parts that are different from sample 1.


Python has a excellent defaultdict datatype from the collections module that is perfect for this application. It is very easy to use:

>>> from collections import defaultdict
>>> x = defaultdict(int)

The essential idea is that you can treat some key as having a default value if it doesn’t exist, so rather than doing this:

>>> x['a'] = x.get('a', 0) + 1

You can do this:

>>> x['a'] += 1

Processing Service Requests

When an instance of the app.Application receives a request it attempts to look up a function based on the message. So when a WhoIsRequest APDU is received, there should be a do_WhoIsRequest function.

The beginning is going to be standard boiler plate function header:

def do_WhoIsRequest(self, apdu):
    """Respond to a Who-Is request."""
    if _debug: SampleApplication._debug("do_WhoIsRequest %r", apdu)

The middle is going to process the data in the request:

# build a key from the source and parameters
key = (str(apdu.pduSource),

# count the times this has been received
whoIsCounter[key] += 1

And the end of the function is going to call back to the standard application processing:

# pass back to the default implementation
BIPSimpleApplication.do_WhoIsRequest(self, apdu)

The do_IAmRequest function is similer:

def do_IAmRequest(self, apdu):
    """Given an I-Am request, cache it."""
    if _debug: SampleApplication._debug("do_IAmRequest %r", apdu)

It uses a diferent key, but counts them the same:

# build a key from the source, just use the instance number
key = (str(apdu.pduSource),

# count the times this has been received
iAmCounter[key] += 1

And has an identical call to the base class:

# pass back to the default implementation
BIPSimpleApplication.do_IAmRequest(self, apdu)

Printing Results

By building the key out of elements in a useful order, it is simple enough to sort the dictionary items and print them out, and being able to unpack the key in the for loop is a nice feature of Python:

print "----- Who Is -----"
for (src, lowlim, hilim), count in sorted(whoIsCounter.items()):
    print "%-20s %8s %8s %4d" % (src, lowlim, hilim, count)

Pairing up the requests and responses can be a useful excersize, but in most cases the I-Am response from a device will be a unicast message directly back to the requestor, so relying on broadcast traffic to analyze device and address binding is not as useful as it used to be.