472 results about "Ternary content addressable memory" patented technology

The present invention provides for a reconfigurable packet classifier using CAM. The invention is directed to packet classification for switching/routing systems where the router's system resources are limited and the customer requirements from the router are variable. The invention addresses the CAM constraint (e.g. search key width) problems of CAM-based classification systems, by allowing a reconfigurable selection of packet fields and/or payload bits to be used in the definition of the search key. For any given incoming packet, a subset of that incoming packet may be statically chosen to fit that particular CAM architecture and to create a particular CAM search key. This provides router deployment flexibility within networks and, thus, cuts costs.

Disclosed is a method and system for multi-character multi-pattern pattern matching. In the multi-character multi-pattern pattern matching method, patterns in an input stream are detected by transitioning between states of a “compressed deterministic finite state automaton (DFA)”, with each transition based on multiple characters of the input stream. The compressed DFA is created by compressing an original DFA, such as an Aho-Corasick DFA, such that each state of the compressed DFA represents multiple consecutive states of the original DFA and each transition between the states of the compressed DFA is a combination of all of the transitions between the multiple consecutive states of the original DFA. This method can be implemented using a Ternary Content-Addressable Memory (TCAM) to store the transitions of the compressed DFA and compares the transitions with multiple characters of an input stream at a time to detect patterns in the input stream.

Disclosed is an apparatus and method for distributing traffic to one or more selected cache systems or servers. In general terms, a content addressable memory (e.g., a TCAM) is utilized to distribute traffic among a plurality of cache systems or servers. The content addressable memory is populated with a plurality of entries. Each entry within the content addressable memory generally indicates an action to be performed on a packet, such as to redirect the packet or to forward the packet to its original destination. When the action indicated by the content addressable memory is to redirect the packet, the content addressable memory also indicates where (e.g., to which cache system) to redirect the packet.

The number of content addressable memory (CAM) lookups is reduced from two to one. Each side (left and right sides) of a CAM is programmed with network addresses, such as IP addresses, based on certain bits of the network addresses. These bits of the network addresses (which represent packet routes) are examined and used to determine whether the particular network address is to be placed on the left or right sides of the CAM. The grouping of certain network addresses either on the left or right sides of the CAM can be performed by examining an individual bit of each network address, by performing an exclusive OR (XOR) operation on a plurality of bits of each network address, and/or by searching for bit patterns of the network address in a decision table. Network addresses that cannot be readily assigned to a particular side of the CAM using these grouping techniques are programmed into both sides of the CAM. During packet routing, techniques similar to the grouping techniques that populated the CAM are used to determine which of the two sides of the CAM is to be searched.

A method and apparatus for input data selection for content addressable memory. In one embodiment, the apparatus includes an array of CAM cells, a select circuit adapted to generate a plurality of select signals each indicative of a segment of input data provided to the CAM apparatus, and switch circuitry including a plurality of programmable switch circuits each programmable to output a respective bit of the input data as a comparand bit for the array of CAM cells in response to one of the select signals.

A system and method provide for efficient classification of long strings of data, such as network messages. The system, which may be a classification engine for use in a network device, is configured to include one or more stages having one or more banks of ternary content addressable memories (TCAMs). The TCAM banks of each stage are organized into one or more groups and each group processes the network messages for a different set of the network device's ports. The system further includes at least one memory resource that is shared by the TCAM banks of all groups. That is, the parallel banks of TCAMs operate in a distributed fashion while the shared memory resource operates in a centralized fashion. Accordingly, the system can process network messages at high speed while minimizing the number of required components.

A multi-stage (e.g., two-stage) packet-based lookup process using a Ternary Content Addressable Memory (TCAM) divided into partitions. The result of a first stage lookup is used to selectively search one of a plurality of TCAM partitions during the second stage. A subset of destination address bits may be used in the first stage to hash to a selected partition for the second stage. Alternatively, a partitioning algorithm segments a routing trie into partitions, and then, either a separate, small TCAM or one of the partitions is used in the first stage to map a prefix of the destination address to one of the TCAM partitions for use in the second stage. The “trie-based” algorithms may advantageously partition the trie such that each second stage partition comprises a substantially contiguous sequence of routing prefixes in a post-order traversal of the routing trie, together with one or more covering prefixes thereof.

A method of managing rule entries of a Ternary Content Addressable Memory (TCAM) in a routing system includes: creating a hash table having a hash key corresponding to each entry; creating a single linked list for linking nodes, the single linked list using the entry of the hash table as a head node and including rule IDs and sequence IDs assigned according to a rule input order; and a double linked list having an independent head node, the double linked list bidirectionally linking the nodes constituting the single linked list according to an order of the sequence IDs. Thus, the packet classifying/filtering rule can be easily added to the TCAM or deleted from the TCAM only with minimal information. Also, the sequence ID reassignment process, required for storing as many rules in the TCAM as possible according to the priority of the rules, is performed when a certain time elapses following rule addition or deletion, thereby reducing a latency that may be caused upon setting the packet classifying/filtering rule.

The invention discloses a low-power ternary CAM by utilizing four encoded comparand datalines, C0, C1, C2, and C3 in a twin ternary cell. The twin ternary cell is a composite of two ternary CAM bits. The two binary CAM bits are coded so that only one of four comparand datalines is toggled during a compare operation. The encoded data is stored and used for comparison. In one embodiment, the four possible states for the 2 bit comparands are coded as 0001, 0010, 0100, and 1000.

A method of handling a datagram in a network device is disclosed. The steps include receiving a datagram, with the datagram having multiple field values, at a port of a network device, parsing the received datagram to obtain the field values, applying the parsed field values to a Ternary Content Addressable Memory (TCAM), determining matches between the parsed field values and predetermined criteria in the TCAM, indexing into a policy table based on the determined matches to obtain an action entry and taking an action based on the obtained action entry.

Methods, systems, and computer readable storage medium embodiments for configuring a lookup table for a network device are disclosed. Aspects in these embodiments include generating a decision tree based upon bit representations of respective data entries from a plurality of data entries where one or more of the plurality of data entries are represented at respective nodes of the decision tree, storing a first bit pattern corresponding to a selected node from the decision tree in a content addressable memory (CAM) at a location associated with an index, and storing one or more second bit patterns at an address in a second memory. The one or more second hit patterns correspond to the one or more data entries represented at the selected node, and the address is associated with the index. Embodiments also include searching a lookup table in a network device.

A hash-CAM (H-CAM) which may work with a controller and a memory containing a database of either search values and associate content or associate content by itself The H-CAM includes at least one set of paired hash units and CAM units and at least one logic unit. The CAM units hold values known to cause hash collisions in the respectively paired hash units, and the logic unit prioritizes the hash and CAM unit outputs to a single address value usable to access the memory and obtain a search result at the controller that is not the result of a hash collision. The H-CAM may optionally include a search data storage to store the search values, so that they need not be stored in the memory, and a comparator to determine and handle newly determined hash collisions.

A method for writing a magnetic random access memory-based ternary content addressable memory cell comprising a first magnetic tunnel junction being formed from a storage layer, a sense layer having a magnetization direction adjustable relative to the magnetization of the storage layer, and an insulating layer between the storage and sense layers; a sense line coupled with the storage layer; a first field line and a second field line, and the first field line being orthogonal to the second field line; comprising: providing a first write data to said storage layer via the second field line to store a first stored data with a high or low logic state; characterized in that, the method further comprises providing the first write data to said storage layer via the first field line to store the first stored data with a masked logic state.

Lookup techniques are described, which can achieve improvements in energy efficiency, speed, and cost, of IP address lookup, for example, in devices and systems employing ternary content addressable memory (TCAM). The disclosed subject matter describes dividing a route table into several sub-tries with disjoint range boundaries. In addition, the disclosed subject matter describes storing sub-tries of a route table between a TCAM and a faster and less costly memory. The disclosed details enable various refinements and modifications according to system design and tradeoff considerations.

The present disclosure concerns a magnetic random access memory-based ternary content addressable memory cell, comprising a first and second magnetic tunnel junction respectively connected to a first and second straps extending on each side of the first and second magnetic tunnel junctions, respectively; a first and second selection transistors, respectively connected to one extremity of the first and second straps; a first and second current lines; and a conductive line electrically connecting in series the first and second magnetic tunnel junctions at their ends opposed to the ones connecting the first and second straps. The cell disclosed herein has smaller size and can be advantageously used in memory devices having a high cell density array.