Preventing Spam: Today and Tomorrow Zane Bonny Vilaphong Phasiname The Spamsters!
Summary Why Prevent Spam How is Spam Prevented What is Wrong With This Picture? What can we do? List Based Approach Algorithm Based Approach Government Legislation Who Did What and Sources Conclusions
Why Prevent Spam Phishing Scams Red Cross Donation Privacy Many want your personal information Out of control 70 to 100 a day at the average office Costly More than 10 Billion a year.
Why Prevent Spam ANNOYING! Who likes spam in their inbox? Can you totally eliminate spam?
How is Spam Prevented Junk Filter – will decide to delete a message or not based on the content of the message. Safe Senders List – this list defines an as safe or not. Imagine an message that is sent through but is deleted by the spam filter. This filter tells the program that it is safe. Safe Recipients Lists – this list is similar to the senders list but is instead used for large groups of people. Blocked Senders List – this is a list of the people that will be treated as junk whether they pass the filter or not.
How is Spam Prevented Never reply to a spam Don’t click any links in a spam Don’t use your home or business address Preview your messages before you open them Disguise your address
What is Wrong With This Picture? Rely heavily on the user Many of these methods do not provide automatic protection. Lists and filters are rarely used by users Even if they are utilized it takes time to be effective What can we do to help eliminate?
What can we do? More user friendly methods More automatic Handled more on the IT side
List: DNS Black Listing Implementation of an old idea Black list can be formed for an individual This is known as DNS Blacklisting Been in use since 1997 Three requirements for Blacklist Domain Name Server List of addresses
List: DNS Black Listing DNSBL queries First reverses ip Second appends DNSBL with reverse IP Last checks names in list Example IP= DNSBL=bl.black.com Sent to blacklist as bl.black.com Policies vary from blacklist to blacklist What does the list wish to prevent? How do you find the addresses? How long?
List: DNS Black Listing
List: Challenge Response This is an filter in reverse Assumes that all is spam First mail is sent Second challenge is issued to the sender Lastly, if the sender responds then they are white listed
List: Challenge Response A number of problems exist Not all can be responded to Listserv Mailing lists Also what if a spammer used a legitimate address?
List: Bounce Messages What is this? Send one each time a spam is sent A few problems…. Spammers don’t care Forged return address Pretty easy to tell by header if it is real or not
Algorithm: Bayesian Probability Bayesian achieves 98%+ spam detection rate using mathematical approach. How does it work? Uses ham files Ham files contain legitimate . For example: The word “free” can be recognize within the data base files of ham. If the word “free” spell differently the Bayesian filter will detected as spam.
Algorithm: Chung-Kwei Named after Feng-Shui figure This figure was a symbol of protection Chung-Kwei is designed to protect business Part of SpamGuru package made by IBM Uses Teiresias algorithm to discover patterns for spam-vocabulary
Algorithm: Chung-Kwei Spam-vocabulary is what is used to filter s before reaching end user. White can remove spam from the spam-vocabulary. Query method then classifies
Government Legislation Why come up with a fancy technique at all why not just ask Uncle Sam for help? Consider the Do Not Call Registry Fairly effective at deterring telemarketers Legal action is available if the telemarketers do not comply On the flip side…. Legal questions arise And constitutional questions
Who Did What? Vilaphong… Algorithm based approaches Government legislation Conclusion Zane… List based approaches PowerPoint Intro
Sources Boyce, Jim. “What to do with all that spam”. Microsoft. 1 May Nov “DNSBL”. Wikipedia. 13 Oct Nov Gowan, Frith. “Don't Get Lured by Phishing Scams”. Techsoup.org. 12 Dec Nov Orlov, Gregory. “Spam: prevention is better than cure!”. BCS. 1 Jan Nov Rigoutsos, Isidore and Huynh, Tien. “Chung-Kwei: a Pattern-discovery-based System for the Automatic Identification of Unsolicited Messages (SPAM)”. IBM Thomas J Watson Research Center. 1 Jan Nov “Section 7 - Spam Prevention”. SORBS. 1 Jan Nov Stuart, Anne. “Canning Spam”. Inc.com. 1 May Nov Tenby, Susan. “Things You Can Do to Prevent Spam”. Techsoup.org. 12 Nov Nov “Why Bayesian Filtering is the Most Effective Anti-Spam Technology”. GFI.com. 1 Jan Nov
Conclusion Have many prevention methods already implemented Most important improvement that can be made is automation Have listing methods and algorithms. algorithms tend to yield the best results Simple lists were sufficient in past Today Spam has evolved to a point that it requires “smarter” methods to prevent it The prevention of spam will undoubtedly become more of issue in the future and cost business a consumers more money A fool proof prevention is unlikely Only 100% way is Government Regulation That also has drawbacks
Questions?