Lucent Technologies Inc.

At&t Company

Alan J Literati

Richard Alan Windhausen

David Martin Sueper

John Michael Nichols

Edward J Mooney

Joseph S Oliner

An arrangement (FIG. 1) that allows banks (15, 25) of up to eight duplicated active devices (30-37, 40-47) to be connected to a SCSI bus (14) and to share SCSI device IDs, and also allows multiple hosts (10, 20) to share the SCSI bus and the connected active devices. Tri-state buffers (200) act as gates (13, 23, 38, 48) to interface host adapters (12, 24) of a plurality of hosts, and active devices of a plurality of device banks, to the SCSI bus. The gate of only one host adapter and the gates of only one bank of devices are enabled at any one time. Only the devices of the enabled bank can receive and respond to addresses from the SCSI bus, while the devices of the other banks are prevented from receiving and responding to addresses from the SCSI bus. Each host has a tri-state buffer acting as bank selector (16, 26) for selecting one of the device banks, by enabling the gates of the devices of the selected bank through a bank select bus (50). Only one host's bank selector is enabled at any one time.

Arrangement for expanding the device capacity of a bus

A timing control arrangement that dynamically controls the distribution of timing information in a distributed digital communication system having standard nodes and prioritized master nodes. Reference timing signals to control the system are distributed from the highest priority master node currently operational. Should a higher priority master node than the current master node become operational, the higher priority master node automatically takes control of the timing for the system. Each node is connected by links to at least one other node, and each node receives timing signals and a timing priority number (TPN) from all of the connected node. The TPN defines the number of links over which the timing signals travel to the receiving node from the current master node. Each standard node selects timing signals and TPN from the link having the lowest TPN and transmits the selected TPN (after incrementing the selected TPN by one) and timing signals on all connected outgoing links. When each master node becomes operational, the node transmits a TPN equal to M*N+1 where M is the priority number of each master node and N is equal to the number of links in the shortest path between the most distant nodes where distance is measured by counting the number of links between nodes. When a master node receives a TPN less than or equal to M*N, the master node enters the standby state and assumes the functions of a standard node since the received TPN indicates that a higher priority master node is controlling the timing of the system. However, if a standby master node receives on all connected links TPNs equal to a predefined maximum (which indicates that no master node is controlling the system), the standby master node assumes the functions of an active master node and starts to control the timing of the system.

Distributed timing recovery for a distributed communication system

Growing community of inventors

Westminster, CO, United States of America

Alan J Literati