Patent EP0696154A2 successfully refreshed from Google Patents.
EP0696154A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Method for the optimal resource allocation for the transport of variable-bandwith data flows in ATM networks, and cross-connect node utilizing the method
Assignee: Telecom Italia SpA
•
Filing: 1995-07-24
•
Granted: —
•
Expires: 2015-07-24
Abstract
Abstract A flow of ATM cells to be cross-connected performs two consecutive passes through the connection network (PX) of a node. In the first pass the virtual channels of the input virtual paths (A, B, C) are distributed among intermediate virtual channel bundles (A'...C''') each with an overall peak bandwidth equal to that of the virtual path of origin of its contributing channels; between the first and the second pass, the virtual channels are cross-connected and the related cells are temporarily memorized, and at the same time the cell flow to be associated to an output virtual path is shaped; and in the second pass the virtual paths are cross-connected, Temporary memorization and shaping are accomplished in a device (SMS) connected between the outputs and the inputs of the connection network (PX) of the node.
Inventors
- 👤 Antonio Cappellari
- 👤 Paolo Coppo
- 👤 Matteo D'ambrosio
- 👤 Vinicio Vercellone
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-10-29 16:54
CPC Classifications
H; H04; H04L; H04L12/00; H04L12/54; H04L12/56; H04L12/5601; H04L12/5602; H04Q; H04Q11/00; H04Q11/04; H04Q11/0428; H04Q11/0478; H04L2012/5603; H04L2012/5609; H04L2012/561; H04L2012/5619; H04L2012/5629; H04L2012/5631; H04L2012/5632
Claims Analysis
Total Claims: 13
Claim 1
Independent Claim
306 words
Method of cross-connecting variable-bandwidth data flows inside a cross-connect node of an ATM network, in which: data flows are composed of ATM cells transported along the network through virtual channels grouped into virtual paths; a peak bandwidth is assigned to the virtual paths during the set-up of the connection; and in correspondence with each node, traffic is policed at the node input and output, to prevent traffic coming from different virtual paths and directed to the same virtual path outgoing from the node from exceeding the node capacity, and traffic is also shaped, to make the emission of the cells related to a given virtual path as uniform as possible, characterized in that: the flow of ATM cells makes two consecutive passes through the connection network (PX) of the node, such that: in the first pass the virtual channels of the input virtual paths (A, B, C) are distributed among intermediate bundles of virtual channels (A', B', C'; A'', B'', B''', C'', C''') each with an overall peak bandwidth equal to that of the virtual path of origin of the contributing channels; between the first and the second pass, a cross-connection of the virtual channels and a temporary memorization of the related cells are performed, by introducing the cells into service queues associated with the output virtual paths, together with a shaping of the cell flow to be associated with an output virtual path; and in the second pass the virtual paths are cross-connected, to multiplex output virtual paths on the output lines of the node; and in that , for the temporary memorization, messages are accepted only if a threshold for the length of the respective queue has not been exceeded or, if the threshold has been exceeded, only if they are high priority messages, thereby performing a selective discarding process at the message level.
Claim 2
Dependent Claim
214 words
Method as claimed in claim 1, characterized in that, for the selective discarding at the message level a state information is memorized for each channel, indicating three possible channel states, namely: - an idle state, which the channel assumes after the end of a message and while waiting for the first cell of a new message; - an accept state, which the channel assumes on receipt of the first cell of a message, if the length of the queue related to the output virtual path into which the message is to be inserted does not exceed the threshold or if the message is a high priority message, and in which the cells are admitted into the respective queue for further processing, the accept state being maintained until the last cell of the message is received; - a reject state, which the channel assumes on receipt of the first cell of a message, if the length of the queue related to the output virtual path into which the message is to be inserted exceeds said threshold and the message is not a high priority message, and in which the cells are not routed to an output queue for further processing, the reject state also being maintained until the last cell of the message is received.
Claim 3
Dependent Claim
70 words
Method as claimed in claim 1 or 2, characterized in that, to accomplish the shaping, an overall bandwidth that is a multiple of an elementary bandwidth is allocated to each active virtual path on a node output and the cells of the corresponding queue are read according to a frame formed by a number of time slots equal to the number of elementary bandwidths necessary to form the overall bandwidth.
Claim 4
Dependent Claim
48 words
Method as claimed in claim 3, characterized in that the traffic shaping frame is built only once for the respective group of simultaneously active virtual paths and is updated in correspondence with the release of one of the paths or with the set up of a new path.
Claim 5
Dependent Claim
82 words
Method as claimed in any of claims 1 to 4, characterized in that the temporary memorization and shaping are accomplished in a device (SMS) inserted between the outputs and the inputs of the connection network (PX) of the node, which device receives all intermediate bundles of virtual channels and builds all virtual paths going out of the node, and in that the intermediate virtual channel bundles realized in the first cross-connection phase are obtained by cross-connecting the virtual paths entering the node.
Claim 6
Dependent Claim
116 words
Method as claimed in any of claims 1 to 4, characterized in that temporary memorization and shaping are accomplished in a device (SMS) that is connected between the outputs and the inputs of the connection network (PX) of the node and that includes a plurality of units (SMS1, SMS2) each handling a part of the virtual channels going out of the node, and in that, for each such unit, the intermediate virtual channel bundles built in the first cross-connection phase are obtained by cross-connecting both virtual paths and virtual channels, in such a way that all virtual channels that are to be transferred to the output virtual paths served by a unit reach the unit itself.
Claim 7
Dependent Claim
279 words
Cross-connect node for an ATM network in which there are transmitted ATM cells carried along the network through virtual channels grouped into virtual paths, each path being assigned, during a connection set-up phase, a peak bandwidth, and the node including a connection network (PX) and means (SMS) to perform a temporary memorization of the cells to be routed to the output virtual paths and a traffic shaping, to make the emission of the cells related to a given virtual path as uniform as possible, characterized in that the means (SMS) for temporary memorization and traffic shaping are connected between a group of outputs and a group of inputs of the connection network (PX) and are able to: - process ATM cell flows that have already undergone a first cross-connection phase, in the connection network (PX), during which phase the virtual channels of the input virtual paths (A, B, C) are distributed among intermediate virtual channel bundles (A', B', C'; A'', B'', C'', C''') each having an overall peak bandwidth equal to that of the virtual path of origin of the contributing channels; - cross connect the virtual channels to form shaped and policed shaped output virtual paths to be re-inserted into the connection network (PX), where they undergo a second cross-connection phase for being transferred to the node output lines; - perform traffic shaping at the virtual channel level by selectively discarding cells at the message level, whereby messages are accepted for insertion into service queues associated to the output virtual paths only if a length threshold for the respective queue has not been exceeded or, if it has been exceeded, only if the messages characterized have high priority.
Claim 8
Dependent Claim
299 words
Node as claimed in claim 7, characterized in that the shaping means (SMS) include at least one unit comprising: - buffer means (BU) in which service queues associated with the different output paths are organized; - a control unit (UC) of the buffer means for managing such queues, said unit generating, on the basis of the queue length, a state information for each virtual channel indicating the possibility to accept or not to accept a message arriving on the channel itself, and building and managing a frame for outgoing traffic shaping; - a plurality of input elements (IN1...INn), which are as many as node input/output lines are and comprise each: a FIFO memory (IF1) for temporarily memorizing the cell flows that have undergone the first cross-connection phase; a content addressable associative memory (TLB), which memorizes information related to the output virtual path to which a received cell is to be routed and the state information of the virtual channel to which said cell belongs, supplied by the control unit (UC) of the buffer means; and a controller (CIF1) of the FIFO memory (IF1) which, on the basis of the information contained in the associative memory, may control the routing to the buffer means (BU), for insertion into the queues, only of cells belonging to high priority messages or to messages received over channels for which, in correspondence with the first cell of the messages, the state information indicates the possibility of accepting the message; - a number of output elements (OU1...OUn), each associated with an output line and comprising a FIFO memory (FU1) and a FIFO memory controller (CFU1), which controls the cell routing to the output line on the basis of the information about the shaping frame provided by the control unit (UC) of the buffer means (BU).
Claim 9
Dependent Claim
34 words
Node as claimed in claim 8, characterized in that said buffer means (BU) are also arranged to organize queues related to cell flows to be routed to or supplied by the control unit (UC).
Claim 10
Dependent Claim
26 words
Node as claimed in claim 8 or 9, characterized in that the buffer means (BU) is a shared resource and the service queues are virtual queues.
Claim 11
Dependent Claim
43 words
Node as claimed in any one of claims 7 to 10, characterized in that the temporary memorization and shaping means (SMS) comprise a single unit, and the connection network (PX), in the first cross-connection phase, carries out a cross-connection at virtual path level.
Claim 12
Dependent Claim
82 words
Node as claimed in any one of claims 7 to 10, characterized in that the temporary memorization and shaping means (SMS) comprise a plurality of units (SMS1, SMS2) each managing a part of the node input and output virtual paths, and in that the connection network (PX), in the first cross-connection phase, groups the input virtual channels directed to virtual paths managed by distinct units into distinct bundles having an aggregate bandwidth equal to the bandwidth of the virtual path of origin.
Claim 13
Dependent Claim
95 words
Node as claimed in any one of claims 7 to 12, characterized in that the temporary memorization and shaping means (SMS) are able to allocate each active virtual path on a node output an overall bandwidth that is a multiple of an elementary bandwidth and to organize reading the service queues associated with the virtual paths simultaneously active on such output by reading the cells of the corresponding queue according to said frame, which is formed by a number of time slots equal to the number of elementary bandwidths necessary to form the overall bandwidth.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.