Multiplex communications – Channel assignment techniques – Polling
Reexamination Certificate
1997-05-07
2001-03-13
Patel, Ajit (Department: 2732)
Multiplex communications
Channel assignment techniques
Polling
C370S252000
Reexamination Certificate
active
06201816
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to token ring networks.
DESCRIPTION OF THE PRIOR ART
Conventional token ring networks include a hub or concentrator having a number of ports which can be connected to respective end stations such as PCs and the like. Each port is connected internally within the hub to a respective switching unit, the switching units being connected in a ring. Each switching unit is able either to pass signals out through a port to a connected end station and then to pass returning signals from the end station on to the next switching unit or to bypass the port. In a passive hub, the switching units are operated either manually or at a low intelligence level whereas in an active hub, the hub will include a controller which monitors performance of the ring and also the connection of end stations.
Typically, when a new end station is to be connected to the ring, it is first physically connected to a spare port of the hub and then the end station raises phantom which involves generating a DC signal on the line connected to the hub unit, typically at a 5V level, which indicates to the controller within the hub unit that the end station is present and wishes to be incorporated in the ring.
Token ring networks operate at speeds of either 4 Mbps or 16 Mbps. End stations which attempt to insert into a ring at the wrong speed can cause significant disruption. To avoid this type of disruption, manufacturers of token ring hubs have implemented various speed detection circuits to enable the token ring hub to lock end stations running at the wrong speed out of the ring. However, there are a number of different manufacturers of token ring hubs and since a method of performing speed detection does not form part of the IEEE Standard 802.5, several different methods have evolved.
As a method for performing this speed detection is not standardised, there are several different methods used in the marketplace:
1. Passive Hubs—As a result of the fact that these hubs are passive, they do not contain a speed detect mechanism.
2. Active Retiming Concentrators (ARCs)—These hubs actively retime all data passing through their ports to the ring speed at which they are configured.
3. Active Hub (Idles)—These hubs detect the speed of the inserting end station based on the speed of data on the lobe immediately after the end station raises phantom.
4. Active Hub (Lobe Test)—These hubs detect the speed of an end station's lobe test before they raise phantom and attempt to join the ring.
5. Active Hub (Burst Errors)—These hubs make use of an architecturally defined response from the end station to detect ring speed. Upon the end station raising phantom, they fix their output signal at one level, thus simulating a Burst Error. This causes an 802.5 compliant end station to apply burst error correction to the incoming signal, which involves adding transitions to the incoming signal at the appropriate data rate. This burst error corrected signal is then used to determine the speed of the attaching end station.
From the point of view of an end station attempting to detect the correct speed to operate at, these different styles of hubs present a challenge which to date has not been successively solved. There are a number of partial solutions to the end station ring speed detect problem, but none of them successfully insert into all the types of hubs outlined above.
In an attempt to reduce the configuration burden placed on the end user, token ring adaptor cards coupled to the end stations to provide an interface with the ring have been designed to sense automatically the speed of the network and configure themselves accordingly.
For an inserting end station attempting to detect the correct speed to operate at in a token ring network, these different token ring hubs present a challenge. If the end station does not insert at the correct speed it will be locked out of the token ring network by the token ring hub.
In one attempt to overcome this problem, the token ring adaptor card is provided with hardware to detect starting delimiters in token ring frames from upstream end stations in the ring once the end station has applied phantom drive or raised phantom. If a starting delimiter is detected and it is at the same speed as that with which the adaptor card was initialised by the host, then the end station completes the IEEE Standard 802.5 insertion process. If a starting delimiter is detected at the other speed then an error code is returned to an adaptor card driver in the host or end station which re-initialises the adaptor card at the other speed and restarts the insertion process. In this technique, starting delimiters are detected by the adaptor card using a pair of phase lock loops (PLL), one of which attempts to lock to a 4 Mbps signal and the other to a 16 Mbps signal. A successful insertion speed is stored in flash memory and the end station uses this as the first speed to try on the next insertion of the end station to the ring.
This form of automatic speed detection adaptor card is unable to insert into a token ring correctly for all of the token ring hubs currently in commercial use. Whilst this approach always works with passive token ring hubs, some active token ring hubs take a decision during a standard lobe test or as the end station applies phantom drive as to whether or not to allow the end station to insert into the ring. In a standard lobe test, the MAC layer on the adaptor card actively monitors the connection with the hub unit. Some active hub units monitor the speed of idle data generated by the end station during a lobe test, when the end station is physically connected to a port but before it has even applied phantom drive. If this monitoring indicates that the speed of the idle data is incompatible with the speed of the token ring, the controller will not permit the end station to open into the ring upon detecting the phantom drive signal. If an end station is operating at the incorrect speed at this time, then it will normally end up being locked out of the token ring. The end station must then repeat the process in complete ignorance of the reason for the failure and consequently, may never be able to insert into the ring.
A further problem with existing adaptor cards is that the end stations are capable of transmitting token ring frames whilst performing the automatic speed detection operation. This means that if two or more automatic speed detection capable end stations attempt to join an empty ring at the same time, they can mislead each other into believing that the ring is configured at one speed by detecting starting delimiters in token ring frames transmitted from another automatic speed detecting end station. This could very easily be the wrong speed and thus prevent any fixed speed end stations which are configured to operate at the correct speed for the token ring network from joining the network subsequently.
SUMMARY OF THE INVENTION
According to a first aspect of the present invention, a method of inserting an end station into a token ring network in which the end station attempts to open into the ring network in one insertion mode, the end station responding to an insertion error code relating to an insertion event, before the end station applies phantom drive, to change from the one insertion mode to another insertion mode, wherein a change of insertion mode is a change in at least one of end station speed and end station idling speed.
In the present invention, the detection of an initial failure code relating to an insertion event (an insertion error code) before a end station raises phantom causes the end station to change insertion mode, where a change in insertion mode is a change in at least one of end station speed and end station idling speed. An example of an insertion error code is a Lobe Test failure code. This mode of failure can result when a end station configured for one speed attempts to open into a token ring network configured to operate at a different speed and is protected by an active token ring hub. As a result, the end st
Madge Networks Limited
Nguyen Brian
Patel Ajit
Sughrue Mion Zinn Macpeak & Seas, PLLC
LandOfFree
Token ring speed detection does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Token ring speed detection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Token ring speed detection will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2517292