Abstract: An e-mail firewall applies policies to e-mail messages transmitted between a first site and a plurality of second sites. The e-mail firewall includes a plurality of mail transfer relay modules for transferring e-mail messages between the first site and one of the second sites. Policy managers are used to enforce and administer selectable policies. The policies are used to determine security procedures for the transmission and reception of e-mail messages. The e-mail firewall employs signature verification processes to verify signatures in received encrypted e-mail messages. The e-mail firewall is further adapted to employ external servers for verifying signatures. External servers are also used to retrieve data that is employed to encrypt and decrypt e-mail messages received and transmitted by the e-mail firewall, respectively.

Abstract: Secure bulk messaging mechanism in which, roughly described, a sender first encrypts a message once. The message can be decrypted with a message decryption key. These can be symmetric or asymmetric keys. For each recipient, the sender then encrypts the message decryption key with the recipient's public key. The sender then sends the encrypted message and the encrypted message decryption keys to a store-and-forward server. Subsequently, one or more recipients connect to the server and retrieve the encrypted message and the message encryption key that has been encrypted with the recipient's public key. Alternatively, the server can forward these items to each individual recipient. The recipient then decrypts the encrypted message decryption key with the recipient's private key, resulting in an un-encrypted message decryption key. The recipient then decrypts the message using the un-encrypted message decryption key.

Abstract: An e-mail firewall applies policies to e-mail messages between a first site and second sites in accordance with administrator selectable policies. The firewall includes a simple mail transfer protocol relay for causing the e-mail messages to be transmitted between the first site and selected ones of the second sites. Policy managers enforce-administrator selectable policies relative to one or more of encryption and decryption, signature, source/destination, content and viruses.

Abstract: Secure web-based messaging according to a “push” paradigm is augmented by specific, intended recipient authentication. In particular, a document can be sent to a specified, intended recipient through the Web using e-mail recipient notification, and the recipient is authenticated prior to delivering the document to the recipient. Such authentication prevents a cracker from snooping a delivery notification e-mail message and retrieving the document prior to retrieval by the true intended recipient. In addition, such authentication of the recipient is driven by the sender such that prior participation by the recipient in the messaging system according to the present invention is required.

Abstract: A server interacts with a sender to form a package which can include one or more attached data files to be sent to one or more recipients, and the server applies a policy established by a policy authority of the sender to the package. Since the sever both forms the package through interaction with the sender and applies the policy, any violations of the policy by the package can be brought to the sender's attention during an interactive session with the sender and before encryption of all or part of the package. As a result, the sender is educated regarding the policy of the sender's policy authority, and the sender can modify the package immediately to comport with the policy. The server delivers the package to the one or more intended recipients by sending notification to each recipient and including in such notification package identification data, e.g., a URL by which the package can be retrieved.

Abstract: Dynamically filtering and classifying messages, as good messages, bulk periodicals, or spam. A regular expression recognizer, and pre-trained neural networks. The neural networks distinguish “likely good” from “likely spam,” and also operate at a more discriminating level to distinguish among the three categories above. A dynamic whitelist and blacklist; sending addresses are collected when the number of their messages indicates the sender is good or a spammer. A dynamically selected set of regular expressions input to the neural networks.

Abstract: Methods and apparatuses for providing cryptographic assurance based on ranges as to whether a particular data item is on a list. According to one computer-implemented method, the items on the list are sorted and ranges are derived from adjacent pairs of data items on the list. Next, cryptographically manipulated data is generated from the plurality of ranges. At least parts of the cryptographically manipulated data is transmitted onto a network for use in cryptographically demonstrating whether any given data item is on the list. According to another computer-implemented method, a request message is received requesting whether a given data item is on a list of data items. In response, a range is selected that is derived from the pair of data items on the list that define the smallest range that includes the given data item. A response message is transmitted that cryptographically demonstrates whether the first data item is on the list using cryptographically manipulated data derived from the range.

Abstract: An e-mail firewall (105) applies policies to e-mail messages (204) between a first site and a plurality of second sites in accordance with a plurality of administrator selectable policies (216). The firewall comprises a simple mail transfer protocol (SMTP) relay (202) for causing the e-mail messages (204) to be transmitted between the first site and selected ones of the second sites. A plurality of policy managers (216) enforce administrator selectable policies. The policies, such as encryption and decryption policies, comprise at least a first source/destination policy (218), at least a first content policy (202) and at least a first virus policy (224). The policies are characterized by a plurality of administrator selectable criteria (310), a plurality of administrator selectable exceptions (312) to the criteria and a plurality of administrator selectable actions (314, 316, 322) associated with the criteria and exceptions.