PineApp Anti IP Blacklisting Whitepaper 2011
Overview ISPs outbound SMTP Services Individual SMTP relay, not server based (no specific protection solutions are stated between the sender and the ISP backbone) Unlike most business organizations networks, taking protective measures for both inbound and outbound mail traffic, ISP s tend to have no or very little enforcement measures over their end users SMTP outbound traffic. Whether it s from financial or any other considerations, ISP s outgoing mail traffic leaves the ISP s backbone unregulated and with zero inspections. In some occasions, ISP s outgoing SMTP traffic s regulation and enforcement is trusted by the hands of the abuse department. The abuse department deals with user complaints (both from inside and outside of the organization), warns potential spammers to cease their criminal activity, in addition to blocking subscribers who ignored their first warning. In an age of automated, Zombie driven, massive Spam delivery, manual and individual treatments are obsolete and inefficient. DHCP Based IP allocation Most ISP s are allocating IP addresses automatically for their endusers, using dynamic DHCP configuration. That means, that each time an end users connects to the internet, he is being allocated with a new randomly assigned IP address. Upon disconnecting, the IP address comes back to the ISP s DHCP pool, set to be available for usage by another, different end user. Considering the way that RBL s are working, this fact turns out to be extremely crucial for one ISP s blacklisting problem. Decentralization unlike business organizations email service, ISP end users are not limited by domain or referring to one central outgoing email server. ISP s outgoing emails come from multiple domains, not all of which are controlled or directly supervised by the ISP. ISP s mailboxes are not commonly and frequently used as they used to be, whereas most endusers are using Gmail, Yahoo and other webmail services. The fact that sender s destinations, as well as email addresses are different states yet another enforcement challenge for SMTP outbound traffic. In order to perform overall outbound SMTP inspections and enforcement of any sort, it is crucial to find a certain common denominator between all connections. Challenge: Zombies and blacklists. Zombies 25% of all worlds computers are a part of a botnets. Spammers and malicious content distributors see a very little use of spending money over dedicated, easily tracked Spamming servers. Therefore, they have developed new methods of recruiting unknowing endusers computers, in order to assemble their own private Spam distribution army. Customers computers are being used to massively distribute Spam. The internet is filled with mischievous, malware infested websites, masquerading within innocent looking files.
once an Trojan horse contained executable file is being run on the customer s computer, a hidden exploit is being opened, set to massively distribute Spam (or other types of malware) while exhausting the system resources of the unknowing end user s computer, as well as burning end user s current IP address clean reputation. Spamcop/Spamhaus blacklists almost all ISP s worldwide are subscribed to at least one RBL service. RBL s (Real Time Blackhole Lists) are voluntarily Anti Spam bodies, in charge of classifying and blocking IP addresses globally, on account of Spam and other malicious content s email distribution. Whenever messages are leaving the ISP backbone by the numbers, and carrying malicious content from a single source, they are being classified and blacklisted accordingly. RBL s classification methods: a. Honeypots Honeypots are mailboxes, set to capture spam and analyze it accordingly. These mailboxes are not directly affiliated with any real persons or genuine mail recipients, so any email arriving to them is probably Spam, advertising junk email or any other type of illegitimate content. Location of email sender is done by checking the message s headers for the source IP address. The RBL s admins are locating the ISP which relayed the message s delivery, and delivering a warning message for the Spam deliverer s abuse department. second Spam conviction results in an introduction of the criminal IP address to the RBL s blacklist. Thus, the offending IP address will be blocked from delivering emails to all email servers subscribed on the RBL s service. b. Reporting/abuse tools another way of blacklisting IP addresses is by receiving complaints from the service providers abused department, which are subscribed to the RBL s service. The abuse representative delivers a complaint email message, containing both message content and message s headers (containing the source IP address from which the message has originated), and according to customer s complain and number of Spam convictions for this IP address it is being stated within the RBL s blacklist. One can conclude that Spam doesn t have to be necessarily delivered with massive amounts in order for the sending IP address to be blocked. It is merely enough to be convicted in Spam delivery twice and your IP address will be stated in several blacklists. One ought to pay serious consideration to that fact when implementing an Anti blacklisting solution. c. Spam/overall traffic volume classification Another method of Spam calculation is by classifying mail servers according to their risk level. According to this method, every source IP address is a mail server. Each Mail server/ip address is calculated for its overall traffic within a timely interval of 30
days. If Spam volume is found to be significantly high, this IP address will be classified as risky and will be blacklisted. As shown above, the fact that Spammers are using end users innocence and unawareness to tarnish an ISP clean IP address reputation is extremely challenging. User s anonymity and dynamic DHCP IP allocation method returns potentially blacklisted IP addresses back to the ISP s pool upon disconnection. Thus, the situation may occur: 1. Zombie infested computers can receive a new, clean IP address upon disconnecting and reconnecting, and can resume their Spam delivery (canceling the blacklisting effect). 2. Other customers may receive a blacklisted IP address, and being banned from email delivery for Spam crimes they did not commit Objective: When deploying anti blacklisting array on ISP networks, one ought to take high importance for several key aspects: Major objectives 1. Locating and stopping Spammers prior to completing delivery phase a. IP rate limiting system massive delivery from a legitimate, business based mail server is normal. However, massive email delivery from an end user s home workstation points on a problematic behavior, and probably Spam delivery attempt. Therefore, it is important to limit the amounts of connections, messages or SMTP sessions per one IP address down to a probable amount. Connections should be sufficient to assure proper functioning, in addition to preventing IP blacklisting according to Spam traffic volume calculation. b. Content scanning As shown above, Spam blockers and RBL s are not only working according to delivery amounts. Limiting rates may narrow down blacklisting in one ISP, but still with insufficient size. Message content scanning ought to take place, in order to lower blacklisting down to satisfactory levels. The major problem with content analysis for ISP scale traffic is that it requires a hefty load of system and network resources, in order to be performed properly. As mail content scanning is yet crucial, it cannot be given up. Therefore, content inspections ought to take place, but to be as less consuming in both system and network resources as possible.
2. Creating IP transparency national regulatory compliance in several geographic locations worldwide are requiring full transparency for SMTP relay arrays, in order not to harm RBL s proper work. If an SMTP relay solution is untransparently stated between the TSP s backbone and the TSP s gateway, it may serve an opposite cause of the one he s stated for: ALL emails will be listed as originated from one source, and, since it s a nation scale outgoing SMTP traffic, it will blacklisted within seconds (regardless of the blacklisting method taking effect). Solution PineApp anti blacklisting array Based on Mail SeCure s proven defensive capabilities, PineApp Anti blacklisting array is set to provide a powerful, comprehensive answer to outbound SMTP blacklisting issues within an ISP network. IP rate limiting If any attack or email blast occurs from a known, unauthorized IP address, it is automatically blacklisted according to traffic volumes. The system allows you to limit maximum messages and sessions per IP within Day/Hour/Minute (Default unlimited). Internal IP Reputation system This highly efficient engine uses proprietary technology to detect and block Spam originated from Zombies within the ISP s pool at the SMTP session level. Thus, enforcement is done internally, and the risk of blacklisting as a result of mass delivery volumes is reduced significantly. Commtouch RPD technology When checked, the Commtouch Recurrent Pattern Detection (RPD ) technology database lookups are activated (default checked). RPD technology analyzes large volumes of email traffic in real time, and is able to detect new Spam and Malware outbreaks as soon as they emerge, as well as mail sent from Zombies (language independent!), according to a repeating pattern within the actual mail message. Deep inspection Engine When checked, PineApp s Anti Spam built in Heuristic and Bayesian engines are activated. Traditional, context based inspections are providing another layer of content inspections. Working under transparent connectivity scheme PineApp Anti blacklisting array provides full transparency during the entire inspection process. Traffic is redirected in to the array, using routing rules, where it is being fully inspected without leaving any foot stamp or any record within the message s headers and without changing the source IP address by any means. Thus, the entire inspection process remains fully transparent.
Benefits 1. Rapid ROI the reliance on external databases, as well as the minimized yet highly efficient content inspection methods, are leading to a sharp reduction in system resources consumption, and as a result a major financial savings in maintenance and infrastructure related expenses. 2. Statistical analysis PineApp anti blacklisting array is supplied with a robust set of textual and graphical statistic sheets, providing tier based statistical data. The anti blacklisting statistics section provides a better understanding and a measure for the array s level of efficiency, as well as traffic trends identification. 3. Thanks to the array s tough, internal IP enforcement and regulation, ISP s IP addresses are regaining credibility and getting cleaned. 4. Logging and auditing features PineApp Anti blacklisting array supplies peripheral logging system, providing a real time tracking mechanism for all SMTP session s phases (including content inspection), in a highly detailed manner. Legal holds allow reversibility, and the releasing of messages that were accidently blocked, in an adjustable time frame, according to customer s needs. In addition, Syslog exporting is fully supported 5. Regulation compliance PineApp s Anti blacklisting array s transparent work scheme, in addition to the Mail SeCure based logging & statistics modules, are improving the ISP s content compliance and are helping in achieving maximized regulation. 6. Optional IP whitelisting the anti blacklisting array supplies tools for IP based whitelisting, set to skip rate limit based inspections for certain IP addresses (efficient for legitimate mailing lists, for example)