SMTP Server Logs
All email servers have the ability to maintain logging information. If available, these logs are usually a more reliable source of information than the mail headers. Because of the high volume of email traffic, the ISP may not store records for very long. SMTP server logs are important tools, used by both ISPs and user organizations to troubleshoot email problems and track down the unauthorized use of mail. The only way to completely verify the path of the mail is for the administrator of each host to check the logs to find that they both sent the message to the recipient as indicated and that their host received it from a specific IP address. If you are lucky, you can follow the chain all the way back to an original IP address. Examination of the original system, which might need to be cross-referenced against RADIUS logs and phone records, is the only way to prove that a specific message existed at a spe-cific place. Accessing this workstation usually requires a court order, but once a perpetrator is confronted with having the incriminating evidence on his or her personal computer, a confession is often not far behind.
Usenet is a vast, distributed bulletin board, consisting of thousands of hierarchically arranged topics contributed to and read by millions of people worldwide. It was originally developed during a time in which Unix computers used modems for external connectivity. Email and even individual files could be sent between noncontiguous UNIX computers, as long as you knew the intervening series of connecting systems and chained their names within the destination mail address (such as ober! doebling!seismo!krampus!kirk). Usenet still takes advantage of this same point-to-point model, although now it uses the Internet as a transport backbone. As shown in Figure 2-14, when a post is made, it is transmitted through a chain of news servers and then it is diffused back out through all of the servers in the world that subscribe to the same newsgroup.
Figure 2-14 Example news server hierarchy
Usenet has adapted itself well to changes in Internet technology, resulting in a nearly seamless integration of traditional news readers with email and Web interfaces. Today, Usenet users have their choice of news clients, email or Web interfaces, both for posting and for reading news messages (see Figure 2-15). Some Usenet mail gateways are also configured as email distribution listservers, enabling Internet citizens to have all the postings for a certain newsgroup delivered directly to their mailboxes, either a message at a time or collected periodically into digests. The implications to you as the tracker is that your attempt to research the origin of a news message may involve multiple types of servers.
Figure 2-15 Those reading and posting news have a choice of using a traditional news reader, a web-based service, or an email gateway.
If you are not familiar with Usenet, and have not configured a client to read postings, the best place to start is the original Internet archive for Usenet, the DejaNews archives provided by Google.8 An example search is shown in Figure 2-16.
Figure 2-16 Example Google page
Inappropriate material is probably the most significant use of news that you are likely to encounter in an investigation. Unfortunately, huge amounts of pornographic pictures are disseminated through news, and it is common for some corporate employees to download this material, exposing their employers to the risk of sexual harassment lawsuits. Pictures, and other binaries, will be found in the uuen-code format (see Chapter 4).
Usenet can provide a forum where you can publicly ask technical questions. You may also want to search archived news postings to see if you can find anything related to one of your suspects or an organization involved in your investigation. Some suspects may be active news posters. Searching for their names on a server such as Google may give you valuable clues to their activities and interests. At times, news postings are directly relevant to an investigation. It is not unheard of for disgruntled employees to try to damage the reputation of their company by making news posts containing harmful information. In some cases, it is possible to trace these inappropriate postings to IP addresses actually within the corporation.
Tracking Usenet Posts
Just like when you are tracking email, you are dependent not only upon the cooperation of all the news server administrators relevant to the message that you are tracking, but you are also dependent upon their having adequate logs. News volume tends to be very high, so the information in the logs is extremely volatile. It is not unusual for a busy Usenet site to turn over their logs every day, so if you want to track a posting, your chances of success are slim if you can't do it within 24 hours. The process you follow will be very similar to tracking a mail message, and some ISP abuse department staffers feel that news is easier to trace because, after some practice, the bogus header information is relatively easy to discern. Just like when tracking email, you work your way back from the recipient and verify each of the machine names in the path. Either you'll find a bogus connection point, which is probably where the message was inserted, or you'll actually verify that the message apparently did transit all of the hosts shown. If the origination host has a record of the poster, you can either subpoena that organization for more information or notify their abuse department.
Let's take a look at the header of a news posting. The following one is an example of a pornographic post that has spent quite a lot of time zipping around the Internet:
From: "YourMate" firstname.lastname@example.org Newsgroups: alt.binaries.dominion.erotica, alt.binaries.dominion.erotica.female, alt.binaries.erotica, alt.binaries.erotica.amateur.female, alt.binaries.erotica.blondes, alt.binaries.erotica.centerfolds Subject: - FREE CD !!! Date: Sun, 5 Dec 1999 10:59:38 0500 Lines: 2484 X-Newsreader: Microsoft Outlook Express 4.72.3110.1 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 NNTP-Posting-Host: dialup599.xxx.com Message-ID: email@example.com X-Trace: 5 Dec 1999 10:52:43 -0500, dialup599.nni.com Path:news.rdc1.nj.home.com!newshub2.home.com!newshub1.home.com!news.ho me.com!feeder.via.net!news.idt.net!netnews.com!newspeer1.nac.net!news. newyork.net!ffx.uu.net!uunet!ams.uu.net!tank.news.pipex.net!pipex!news - lond.gip.net!news.gsl.net!gip.net!nntp.news.xara.net! xara.net!gxn.net!news.good.net!news.xxx.com!phila-dialup599.xxx.com Xref: newshub2.home.com alt.binaries.dominion.erotica:30040448 alt.binaries.dominion.erotica.female:30252819 alt.binaries.erotica:31064499 alt.binaries.erotica.amateur.female:30199540 alt.binaries.erotica.blondes:30083027 alt.binaries.erotica.centerfolds:30058010
With the understanding that virtually everything on this header can be faked, each header purportedly contains the following:
From: This is the name and email address of the original poster.
Newsgroups: In this case, the message was cross-posted to six different news-groups.
NNTP-Posting-Host: This is the machine from which the posting was sent. It is usually the same as the lower rightmost entry in the Path header.
Message-ID: This is a unique serial number assigned to every post on NNTP servers. The logfile on news.xxx.com should have an entry that associates this serial number with this specific message and a specific account. This ID should be unique throughout the worldwide Usenet, and it can be used to cancel messages across all (well, most) news servers worldwide.
Path: This is the meat of your investigation. In reverse chronological order, it shows every host that the message has transited. The first host, news.rdc1. nj.home.com, is the host on which this message was received, and it is the only information in this entire header that you can trust. The succeeding hostnames are the hosts that purportedly relayed this message. The next-to-last host, news.xxx.com, is apparently the news server at an ISP, while the final hostname looks like a dial-up account at that same ISP.
Note that on the preceding header, the domain name in the From field is consistent with the domain name in the NNTP-Posting-Host and Path fields. If it were not, you would know immediately that someone was attempting to cover his or her tracks. If they do match, you probably still need to verify the intermediate hosts, but you can start with the origination point. If they don't match, you need to figure out which parts of the path are real and which are bogus. It often helps to obtain other copies of the same message from other servers (either Google or servers belonging to friends of yours) and compare the paths. Whatever part of the paths is consistent among the different news hosts probably contains the actual host at which the message was inserted.
Like mail hosts, Network News Transfer Protocol (NNTP) hosts may or may not accept posts from people outside of their organization. You can test this by telneting to port 119 on that host. If you cannot connect to that host or you receive a message that posts are not accepted, odds are lower that a bogus post was made using that host as the news server. However, if you receive a message that indicates you can connect, it means that other people can connect to that host also. The chances are reduced that the administrator will really know who is using the NNTP server. Such a session may look like this:
C:> telnet ferkel.piglet.com 119 200 NNTP Service Microsoft® Internet Services 5.00.7515 Version: 5.0.7515 Posting Allowed
$ telnet news.isp.com 119 200 mercury2.isp.com Netscape-Collabra/3.52 17222 NNRP ready (posting ok).
But if the host does not accept posts, it may look something like this:
$ telnet nntp.mindspring.com 119 502 You are not in my access file. Goodbye.