Abstract: A fast recovery extended method is used to enhance the performance of TCP fast recovery when multiple segment losses occur within a single round trip time between a TCP sender and a TCP receiver. A TCP fast recovery process is performed by a TCP sender, and then a TCP fast recovery extended process is performed by the TCP sender upon receiving acknowledgement of receipt of new data from a TCP receiver in the TCP fast recovery process. The fast recovery extended process determines, following receipt of the acknowledgement of receipt of new data, an excess number of duplicate acknowledgements based upon a count of consecutive duplicate acknowledgement packets. The fast recovery extended process takes a network packet transmission recovery action based upon the excess number of duplicate acknowledgements, and then stores the excess number of duplicate acknowledgements as a number of duplicate acknowledgements for further use.

Abstract: A plurality of web servers (16, 18, and 20) have a common host name, and their authoritative domain server (24 or 26) responds to requests from a local domain-name server (22) for the network address corresponding to their common host name by making an estimate of the performance costs of adding a further client to each of the web servers and then sending the local domain-name server (22) the network address of the server to which the addition of a further client will result in the least performance cost. The performance cost is defined as the difference in the average number of waiting clients, and it takes into account both the additional response time for existing clients and the projected response time for the prospective new client.

Abstract: To determine the name that specifies the Systems Network Architecture logical-unit-characteristics set that a host (12) systems services control point is to associate with a TN3270 server (18) logical-unit process that is to represent a client terminal (14), the server (18) generates the name from the client's device type algorithmically, i.e., without looking the name up in a server-maintained table entered by using the device type as the key. This eliminates the need for server updates when new device types are introduced. Additionally, the algorithm used to generate the name depends on whether the client has negotiated to receive a copy of the data transmitted in the BIND message by which another logical unit initiates a session with the logical unit that will represent the client. In that way, the host's systems services control point can send the client data streams in the form that it expects, and the server is relieved of the need to perform any translation.

Abstract: When a TN3270 client (14) uses the Transmission Control Protocol/Internet Protocol ("TCP/IP")to initiate contact with a TN3270 server (18) to obtain access to an applications program running on a host (12) with which the server (18) employs the Systems Network Architecture ("SNA")protocol to communicate, the server's SNA physical-unit process sends the host's systems service control point an SNA NMVT message that includes a physical-location subfield in a product-set-attributes subvector of a product-set-ID vector. The physical-location subfield includes the client's IP address and TCP port number so that host management applications can determine individual clients' usages without interrupting communications with the server.