The document discusses spam over internet telephony (SPIT) and presents a reference model for prevention systems, classifying various methods to minimize unwarranted calls while maximizing legitimate connections. It outlines numerous strategies including greylisting, reputation systems, and consent-based communication, among others, highlighting a modular design for multi-stage prevention. The authors aim to assist newcomers in understanding and implementing effective SPIT prevention measures.

1. On Spam over Internet Telephony
(SPIT)
Prevention

2. 
SPIT is defined as the transmission of unsolicited
calls over Internet telephony

Using Internet telephony, these costs are
substantially lower

presents a reference model for SPIT prevention
systems and provides a classification for a large
set of available prevention methods

3. 
minimize the probability of blocking legitimate
call

maximize the probability of blocking SPIT calls.

minimize the interactions with the callee

limit the inconvenience caused to the caller that
tries to place a legitimate call

should be general

should be flexible

4. 
No Interactions With
Call

Caller-side Interactions

Callee Interrupted by
Call

Callee Receives Call

Feedback From Callee
After call

5. Lists

simple mechanism

white lists (identities allowed to call)

black lists (identities that should be rejected)

members of white lists and black lists may be
configured

6. Circles of Trust
introducing trusted inter-domain connections
each domain controls its own users and the
domains agree not to send SPIT to each other
can be implemented in SIP by using
authenticated TLS connections between
domains

8. Pattern/Anomaly Detection
detecting suspicious patterns in VoIP traffic to
identify SPIT calls
very general
deciding, based on patterns and signatures
whether the incoming call might be SPIT or
not

9. Greylisting
first call from an unknown user is rejected

10. Computational Puzzle

giving the caller's terminal a resource
consuming task to perform before establishing
the call

attackers use botnets to distribute the cost of
computing puzzles

limits the effectiveness of the solution

11. Sender Check

verify that a caller is a valid sender for the
domain from which he is calling.
Turing Test

conversational method to tell humans and
computers

the judge is a human being

most CAPTCHAs are visual although audio
CAPTCHAs exist as well

12. Consent-based Communication

user A authorize user B, the first time user B tries
to contact user A

it solves the first-contact problem but introduces a
delay until the first call can be placed

14. Content Filtering

blocking email spam is essentially based on
content analysis

cannot be fully applied to prevent SPIT

the receiver has been disturbed by a ringing
phone

15. Reputation System
 attaching reputation score to a contact
 this score can be evaluated based on user
feedback
 it could also be tied to other methods

16. Limited-Use Addresses
 changing the address as soon as the first
spam messages arrive at the address
 new address has to be communicated to all
contacts
 new user has to be able to get the current
address of a recipient

17. Payments At Risk
 charging a fee for the first contact
 refunding that fee if the call was not SPIT
Legal Action
 introducing legislation in all countries to
prohibit the distribution of spam over VoIP
 there will always be countries where it is legal
to send SPIT

18. First-Contact Feedback
 the user can provide a feedback to the server
 an unknown identity is allowed to call exactly
once and then the callee has to provide a
feedback
 there is no first-contact problem

20.  we used a modular design for our prevention
system that flexibly allows linking stages and uses
a modular approach for stages 1 and 2.

22. SPIT PREVENTION AT STAGE 1

All modules examine incoming call signaling and
produce a score

compared to two threshold a low and a high

below the lower threshold, the call is forwarded to
the calee

between the lower and higher thresholds, call is
forwarded to the second stage modulesrejected or
forwarded to a voicemail system

above the higher threshold, either the call is

23. SPIT PREVENTION AT STAGE 2

If the test is successfully passed ,the
dispatcher forwards the call to the original
callee

stage 2 module belonging to the“Turing Test”

based on the assumption that human
conversation follows certain activity patterns

sends a prerecorded greeting message

26. SPIT PREVENTION AT STAGE 5

Software client allows user to terminate the call
• time indicates to the SPIT prevention system
that this was SPIT

white/black list module may add the caller
identity to the black list.

27.  A systematic classification and reference model
for SPIT prevention building blocks
 Has a purpose of helping newcomers and
practitioners in the area to improve their
understanding, and better design and implement
SPIT prevention systems

28.  Juergen Quittek & Savario Niccolini; On Spam
over Internet Telephony(SPIT) Prevention, IEEE
Communication magazine, August 2008
 www.2dix.com

29. THANK YOU