Mailshell The OEM Anti-Spam Leader  
About Solutions Partners Products News Customer Login
 
Partners Testimonials Resellers OEM Case Studies Oracle Collaboration Suite

FAQ

1. How does Mailshell Anti-Spam detect spam?
2. How does Mailshell handle foreign languages?
3. How often are rules updated?
4. How does it prevent false positives?
5. How does it prevent future spam?
6. How reliable is Mailshell's enterprise software?
7. How does it operate?
8. What is the capacity of the Mailshell Network?
9. What is the processing capacity of the Gateway software product.
10. How much is accuracy degraded if we don't connect to the network?
11. What is the performance impact of the software?
12. How do I create and manage per user Spam folders?


1. How does Mailshell Anti-Spam detect spam?

Mailshell looks for, catches and analyzes spam, 24x7, sent to millions of email accounts at AOL, Hotmail, Yahoo!, AT&T, Earthlink, Mailshell, and others representing over 95% of all spam sent out on the Internet.

Mailshell then creates rules based on all attributes of the message to identify and blacklist spam immediately. Our Anti-Spam software automatically checks for new rules and updates itself. Thousands of rules are applied and combined by Bayesian theorems for a final spam probability score. New, never before seen spam is immediately caught.

Mailshell also monitors and maintains the proprietary Mailshell Network database. The Mailshell Network tracks millions of messages in real-time to determine which messages are sent in bulk and whether the bulk messages are wanted by users. The Mailshell Anti-Spam software creates an anonymous fingerprint ID of the email and immediately checks the Mailshell Network if the email is spam. When users approve or block an email, the Mailshell Anti-Spam software immediately updates the Mailshell Network. Spam is blocked throughout the network as soon as it is detected.

Mailshell allows users to customize how spam is caught. The user is allowed to specify a personal 'always approved' list and a personal 'always block' list. The software intelligently adapts based on what the user blocks, allows, reads, replies to, and deletes.

Our rules include major third-party spam databases and RBL blacklists, as well as rules to handle mail bombs, mail DoS, and mail flood detection.


2. How does Mailshell handle foreign languages?

Many of our rules handle foreign languages and the Mailshell Network is language neutral. However, some reduction in the percentage of spam caught is likely for email sent in foreign languages.


3. How often are rules updated?

Rules and weightings are updated usually every 10 minutes.


4. How does it prevent false positives?

False positives in general are due to ambiguity as to the reputation of the bulk sender. We maintain a large database of reputable senders and track user feedback to accurately determine bulk sender reputation. The probability algorithms only catch messages which are very likely spam and very unlikely to be legitimate messages. To learn more about Mailshell Anti-Spam's accuracy versus error rates, see our white paper.


5. How does it prevent future spam?

The software learns and adapts in real-time and personalizes the algorithms for each user so that spammers can not easily target a large audience.


6. How reliable is Mailshell's enterprise software?

The software is based on more than 3 years of development and testing. Over 700,000 users have trusted the Mailshell commercial service for more than 2 years. The software requires very low memory usage and large enterprises can choose even higher availability via clustering support.


7. How does it operate?

Our Gateway proxy software currently requires Linux operating system 2.2+ running on the x86 platform. There is also a version available for Windows 2000/XP. It runs as a SMTP proxy which listens on a specified port and forwards messages to another destination SMTP server.


8. What is the capacity of the Mailshell Network?

It is designed to scale to over 1 billion requests per day.


9. What is the processing capacity of the Gateway software product.

The Mailshell Anti-Spam proxy software currently processes 700,000 messages per day per CPU.


10. How much is accuracy degraded if we don't connect to the network?

The Mailshell Network separates Mailshell's solution from the myriad of products that rely exclusively on rules-based filtering. While the software still catches the vast majority of spam without checking the Network, eliminating this step will likely result in roughly 20% less caught spam. To learn more about Mailshell Network's Network's impact on filtering accuracy, see our white paper.


11. What is the performance impact of the software?

On average, messages are delayed by less than a second. Most of the delay is due to accessing the Mailshell Network database. 99% of messages are delivered within 5 seconds.

12. How do I create and manage per user Spam folders?

To create and manage per user Spam folders, Mailshell currently offers plugins for the following mail servers:

 

 

Mailshell Inc., 695 Fifth Street, Suite #3, San Francisco, CA 94107, tel: 415.348.8728