Networks Associates Technology, Inc.

Wonderware Corporation

Roger Knobbe

Stephen Schwab

Andrew Kyle Purtell

Steven Wong

Two or more computers acting as firewalls share network state data to enhance throughput performance. A firewall creates a separate common TCP control block (CCB) for each group of TCP connections through the firewall having common endpoints. The CCB is a shared data structure comprising a single microstate shared across the group of TCP connections. Each such individual TCP connection has a TCP control block, which instead of a microstate, contains a pointer to the appropriate CCB. Preferably, each firewall receives CCBs from its peers and stores them. Each firewall preferably adjusts data traffic passing through it based on the CCBs stored within it. By adjusting traffic to reduce or eliminate congestion, throughput is enhanced.

System for sharing network state to enhance network throughput

A device that passively monitors arriving and departing data packets on one or more networks, correlates arriving data packets with departing data packets, and calculates a latency estimate based on the confidence of the correlation. The device detects and copies data packets arriving at a network device and the data packets departing from the same network device. A timestamp is stored for each arriving or departing data packet. Latency across a network device can be determined based on the timestamps for correlating data packets. Additionally, latency across a network device per protocol layer can also be calculated. Varying levels of confidence of a latency estimation depend on the operation necessarily performed on the data packet by the network device and the protocol level at which correlation between the arriving and departing data packets can be achieved.

Method and apparatus for passively calculating latency for a network appliance

A system and method for regulating data transmission between a source node and a destination node is disclosed herein. The present method contemplates monitoring system performance so as to detect when the system enters an overloaded state. Such detection may be effected by, for example, determining the extent of the time interval during which instructions are retained in a queue at the destination node awaiting execution. When this interval exceeds a maximum queuing time, the system is deemed to be in an overloaded state. Upon entry into an overloaded state, data transmission from the source node is suspended pending recovery of the system from the overloaded state. In a preferred implementation, overload probe messages are periodically inserted into a message queue at the destination node. When such an overload probe message is determined to have been retained in the message queue in excess of the maximum queuing time, the system is deemed to be in an overloaded state and data transmission is suspended as described above. If it is subsequently determined that an overload probe message has been retained in the message queue for less than the maximum queuing time, data transmission from the source node is resumed.

Growing community of inventors

Torrance, CA, United States of America

Roger Knobbe