Associative memory cells configured to selectively produce binary or ternary content-addressable memory lookup results

Description

TECHNICAL FIELD

One embodiment of the invention relates to communications and computer systems, especially reliable content-addressable memory devices and use thereof; and more particularly, one embodiment relates to associative memory cells configured to selectively produce binary or ternary content-addressable memory lookup results.

BACKGROUND

The communications industry is rapidly changing to adjust to emerging technologies and ever increasing customer demand. This customer demand for new applications and increased performance of existing applications is driving communications network and system providers to employ networks and systems having greater speed and capacity (e.g., greater bandwidth). In trying to achieve these goals, a common approach taken by many communications providers is to use packet switching technology. Increasingly, public and private communications networks are being built and expanded using various packet technologies, such as Internet Protocol (IP). Note, nothing described or referenced in this document is admitted as prior art to this application unless explicitly so stated.

A network device, such as a switch or router, typically receives, processes, and forwards or discards a packet based on one or more criteria, including the type of protocol used by the packet, addresses of the packet (e.g., source, destination, group), and type or quality of service requested. Additionally, one or more security operations are typically performed on each packet. But before these operations can be performed, a packet classification operation must typically be performed on the packet.

Packet classification as required for, inter alia, access control lists (ACLs) and forwarding decisions, is a demanding part of switch and router design. The packet classification of a received packet is increasingly becoming more difficult due to ever increasing packet rates and number of packet classifications. For example, ACLs require matching packets on a subset of fields of the packet flow label, with the semantics of a sequential search through the ACL rules. IP forwarding requires a longest prefix match.

Known approaches of packet classification include using custom application-specific integrated circuits (ASICs), custom circuitry, software or firmware controlled processors, and associative memories, including, but not limited to binary content-addressable memories (binary CAMs) and ternary content-addressable memories (ternary CAMs or TCAMs). Each entry of a binary CAM typically includes a value for matching against, while each TCAM entry typically includes a value and a mask. The associative memory compares a lookup word against all of the entries in parallel, and typically generates an indication of the highest priority entry that matches the lookup word. An entry matches the lookup word in a binary CAM if the lookup word and the entry value are identical, while an entry matches the lookup word in a TCAM if the lookup word and the entry value are identical in the bits that are not indicated by the mask as being irrelevant to the comparison operations.

FIGS. 1A-B illustrate one implementation of a prior art binary content-addressable memory entry 100 and bit cells 101A-B therein. Each bit cell 101A-B includes a storage 105A-B for storing a value to be matched against, which are stored by exerting the read/write line 110 and based on the bit values and their complements on lines 111A-B and 112A-B. In order to determine if binary content-addressable memory entry 100 matches a lookup word, match line 132 is precharged to Vdd based on precharging signal 120 during a precharging phase of the lookup operation. The bit values (and their complements) of the lookup word are applied on bit lines 111A-B and 112A-B, which are compared by comparison circuits 130 during a matching phase of the lookup operation to the stored bit values (105A-B). If any of the entry bit cells 101A-B determine that they do not match, then match line 132 is discharged by the corresponding comparison circuit(s) 130.

This is shown in more depth in FIG. 1B by binary CAM cell 140 with discharge paths 141. The particular one of the two discharge paths 141 that will discharge match line 132 in response to an input bit value not matching the stored value depends on the stored value to be matched, the input bit's lookup value, and the implementation of the binary CAM cell.

Any drop in the voltage of match line 132 is amplified, with any slight change in the voltage of match line 132 quickly detected by match line sense amplifier 135, which provides the corresponding hit/miss indication 136 based on the results of the comparison operations.

FIGS. 1C-D illustrate one implementation of a prior art ternary content-addressable memory entry 150 and bit cells 151A-B therein. The operation of ternary CAM entry 150 is very similar to that of binary CAM entry (FIG. 1A). Each TCAM cell A-B stores a value to be matched against, with this value representing one of the three values: one, zero and don't care/wild card (i.e., match either one or zero). Therefore, a two-bit storage 155A-B is typically used in TCAM cells 151A-B to store its value. Any drop in the voltage of match line 182 caused by comparison circuits 180 is amplified, with any slight change in the voltage of match line 182 quickly detected by match line sense amplifier 185, which provides the corresponding hit/miss indication 186 based on the results of the comparison operations.

For example, in a TCAM cell, a stored value of one might be encoded as bits ‘10’, a zero might be encoded as bits ‘01’, and don't care (often represented as ‘*’) might be encoded as bits ‘00’. The encoding of the bits to be matched are supplied via data lines 161A-B and 162A-B. With this encoding scheme for storing the values in TCAM cells 151A-B and the NMOS implementation shown in more detail in FIG. 1D, an encoding of the lookup value of one or zero for each TCAM cell is just such that if the supplied lookup bit value does not match the TCAM cell 190, then one of the two discharge paths 191 will be turned on to pull down match line 182. For example, with a stored value of one (‘10’), a supplied lookup value of zero (‘10’) is provided such that two transistors will be enabled to form one of the discharge paths 191, causing match line 182 to drop. This voltage drop is detected by match line sense amplifier 185 and a miss signal 186 generated to signal that the stored value in the TCAM entry does not match the lookup word. Note, if the encoding of a don't care value is ‘00’ and each bit is applied to a transistor in a different discharge path, then a discharge path will never become active, and therefore, the TCAM cell will effectively match the value of the corresponding bit of the lookup word.

Associative memories are very useful in performing packet classification operations. As with most any system, errors can occur. For example, array parity errors can occur in certain content-addressable memories as a result of failure-in-time errors which are typical of semiconductor devices.

When a packet classification lookup operation is performed on an associative memory with corrupted entries, a bit error in an entry can result in a false hit, or a false miss. A false hit occurs when the corrupted value of an entry matches the lookup value when it otherwise would not match that entry (and thus another entry or no entry should have been matched). A false miss occurs when an entry should have been matched except for the corruption in the entry. This could result in no entry being matched or another lower-priority entry being matched. When these lookup operations are used for packet classification, an incorrect match or miss presents a problem especially when identifying a route or performing a security classification.

Error-correcting and error-detecting codes are well-known. For example, ANDREW S. TANENBAUM, COMPUTER NETWORKS, Prentice-Hall, 1981, pp. 125-132, discusses error-correcting and error-detecting codes, and is hereby incorporated by reference. Assume a codeword contains n bits of which m are data bits and r are error-correcting or error-detecting bits (e.g., redundant or check bits), with n=m+r. There are many well-known ways to generate the error-detecting and error-correcting bits. Given two codewords, it is possible to determine how many bits differ (e.g., by exclusively-OR'ing or one bit summing the corresponding bits of the two codewords and summing these results). The number of bit positions in which two codewords or a set of codewords differ is called the Hamming distance. A Hamming distance of d, means that it will require d single-bit errors to convert one codeword to another codeword. To detect j errors, a Hamming distance of j+1 is required because with such a code, there is no way that j single-bit errors can change a valid codeword into another valid codeword. Similarly, to correct j errors, a distance 2j+1 code because that way the legal codewords are so far apart that even with j changes, the original codeword is still closer than any other codeword, so it can be uniquely determined.

A prior approach protects the associative memory entries with error detection or correction values when the associative memory is not being used to perform a lookup operation. For example, using a background operation, the associative memory entries are periodically checked and corrected for errors (e.g., read from their location and if an error, the correct value is written back). Another prior approach is to periodically over write each associative memory entry with the correct value. These and other prior approaches do not immediately detect the error, nor detect the error when a lookup operation is performed on the corrupted entry. Thus, there can be significant periods (e.g., several seconds to minutes which can be a very long time in the context of a packet switch) before such corrupted entry is corrected.

Some random access memory (RAM) add error-correcting or error-detecting codes to each memory cell. As part of a read operation of a memory location, the data portion and the error-correcting or error-detecting code is read, which is then used to detect a possible error and/or correct a discovered error in the data portion. This is especially convenient to do as only one set of error-detecting/error correcting circuitry is required (i.e., to operate on the data read from the specified memory location). However, this approach is impractical for an associative memory, as each associative memory entry would need this complete circuitry, and the result of the error-corrected operation for each memory location would need to be compared to the lookup word for every lookup operation.

Shoham et al., “Error Protection for Lookup Operations in Content-Addressable Memory Entries,” U.S. patent application Ser. No. 10/879,254, filed Jun. 29, 2004, is hereby incorporated by reference in its entirety. Shoham et al. discloses techniques for populating content-addressable memory entries with values that have a Hamming distance of n bits and for identifying a matching entry that matches in at least k bit positions, where n>k>0. Other techniques not explicitly disclosed in Shoham et al. for identifying such matching entries could be useful.

Routers typically use TCAMs for identifying routing information as they are efficient at determining a longest matching prefix on an IP route, for example; while bridges typically use a CAM for identifying routing information as they are efficient at determining an exact match on a frame address, for example. Of course, a TCAM can perform CAM lookup operations (e.g., by not programming don't care cases and therefore an exact match is required), but each TCAM cell typically requires substantially more space and may consume more power than a standard CAM cell.

SUMMARY

Disclosed are, inter alia, methods, apparatus, data structures, computer-readable media, mechanisms, and means for selectively producing binary or ternary content-addressable memory lookup results. One embodiment includes: multiple associative memory bit cells configured to determine whether or not their associated value matches a respective search value of multiple search values; circuitry configured to selectively apply search values to the associative memory bit cells in a manner such that the associative memory bit cells produce results as being either (a) n binary content-addressable memory (CAM) cells or (b) m ternary content-addressable memory (TCAM) cells, with n>m>0; wherein said selective application of the search values provides different values to the associative memory bit cells when in the n CAM mode than in the m TCAM mode; and a match detector coupled to the plurality of associative memory bit cells, the match detector configured to identify whether or not at least one of the associative memory bit cells does not match their respective said search value in order to produce the n CAM cell or m TCAM cell matching result.

In one embodiment, in a TCAM bit mode, the program values associated with the two CAM bit cells are a same value when the TCAM bit represents a don't care value. In one embodiment, in a TCAM bit mode, the program values associated with the two CAM bit cells are different values when the TCAM bit represents a one or a zero value. In one embodiment, n is two and m is one. In one embodiment, error protection is used to enable a correct result when e bit errors occur in the stored values in the associative memory bit cells. In one embodiment, 2e+1 associative memory identification bit cells are used for each row or group of associative memory bit cells in providing the error protection.

BRIEF DESCRIPTION OF THE DRAWINGS

The appended claims set forth the features of the invention with particularity. The invention, together with its advantages, may be best understood from the following detailed description taken in conjunction with the accompanying drawings of which:

FIGS. 1A-D are a block diagram of a prior art content-addressable memory entry and bit cells therein;

FIG. 2 illustrates an observed varying of the match reference voltage of a precharged match line of a content-addressable memory cell in response to the number of content-addressable memory bit cells not matching their respective bits of a lookup word;

FIG. 4A is a block diagram of a content-addressable memory entry of a content-addressable memory according to one embodiment;

FIG. 4B is a flow diagram of a process used in one embodiment for identifying whether or not a content-addressable memory cell should be considered as matching;

FIG. 5 is a block diagram of a content-addressable memory used in one embodiment;

FIG. 6 is a flow diagram of a process used in one embodiment for programming TCAM entries;

FIGS. 7A-C are block diagrams of an associative memory bit cell according to one or more embodiments;

FIG. 7D illustrates programming of associative memory bit cells according to one embodiment;

FIG. 7E is a flow diagram of a process used in one embodiment for identifying whether or not an associative memory bit cell should be considered as matching;

FIG. 8A is a block diagram of a row of associative memory bit cells used in one embodiment;

FIG. 8B is a flow diagram of a process used in one embodiment for identifying whether or not a row of associative memory bit cells should be considered as matching;

FIG. 8C is a block diagram of multiple rows of associative memory bit cells used in one embodiment;

FIG. 8D is a block diagram of a group of associative memory bit cells used in one embodiment;

FIG. 9A is a block diagram of multiple rows of associative memory bit cells with each row having multiple associative memory identification cells as used in one embodiment; and

FIG. 9B is a block diagram of multiple associative memory bit cells, with each row having a dedicated associative memory identification cell and one or more associative memory bit cells which can be programmed as associative memory identification cells in accordance with one embodiment.

DETAILED DESCRIPTION

Disclosed are, inter alia, methods, apparatus, data structures, computer-readable media, mechanisms, and means for selectively producing binary or ternary content-addressable memory lookup results.

Note, embodiments described herein include various elements and limitations, with no one element or limitation contemplated as being a critical element or limitation. Each of the claims individually recites an aspect of the invention in its entirety. Moreover, some embodiments described may include, but are not limited to, inter alia, systems, networks, integrated circuit chips, embedded processors, ASICs, methods, and computer-readable media containing instructions. One or multiple systems, devices, components, etc. may comprise one or more embodiments, which may include some elements or limitations of a claim being performed by the same or different systems, devices, components, etc. The embodiments described hereinafter embody various aspects and configurations within the scope and spirit of the invention, with the figures illustrating exemplary and non-limiting configurations. Note, computer-readable media and means for performing methods and processing block operations are disclosed and are in keeping with the extensible scope and spirit of the invention.

As used herein, the term “packet” refers to packets of all types or any other units of information or data, including, but not limited to, fixed length cells and variable length packets, each of which may or may not be divisible into smaller packets or cells. The term “packet” as used herein also refers to both the packet itself or a packet indication, such as, but not limited to all or part of a packet or packet header, a data structure value, pointer or index, or any other part or direct or indirect identification of a packet or information associated therewith. For example, often times a router operates on one or more fields of a packet, especially the header, so the body of the packet is often stored in a separate memory while the packet header is manipulated, and based on the results of the processing of the packet (i.e., the packet header in this example), the entire packet is forwarded or dropped, etc. Additionally, these packets may contain one or more types of information, including, but not limited to, voice, data, video, and audio information. The term “item” is used generically herein to refer to a packet or any other unit or piece of information or data, a device, component, element, or any other entity. The phrases “processing a packet” and “packet processing” typically refer to performing some steps or actions based on the packet contents (e.g., packet header or other fields), and such steps or action may or may not include modifying, storing, dropping, and/or forwarding the packet and/or associated data.

The term “system” is used generically herein to describe any number of components, elements, sub-systems, devices, packet switch elements, packet switches, routers, networks, computer and/or communication devices or mechanisms, or combinations of components thereof. The term “computer” is used generically herein to describe any number of computers, including, but not limited to personal computers, embedded processing elements and systems, control logic, ASICs, chips, workstations, mainframes, etc. The term “processing element” is used generically herein to describe any type of processing mechanism or device, such as a processor, ASIC, field programmable gate array, computer, etc. The term “device” is used generically herein to describe any type of mechanism, including a computer or system or component thereof. The terms “task” and “process” are used generically herein to describe any type of running program, including, but not limited to a computer process, task, thread, executing application, operating system, user process, device driver, native code, machine or other language, etc., and can be interactive and/or non-interactive, executing locally and/or remotely, executing in foreground and/or background, executing in the user and/or operating system address spaces, a routine of a library and/or standalone application, and is not limited to any particular memory partitioning technique. The steps, connections, and processing of signals and information illustrated in the figures, including, but not limited to any block and flow diagrams and message sequence charts, may typically be performed in the same or in a different serial or parallel ordering and/or by different components and/or processes, threads, etc., and/or over different connections and be combined with other functions in other embodiments, unless this disables the embodiment or a sequence is explicitly or implicitly required (e.g., for a sequence of read the value, process the value—the value must be obtained prior to processing it, although some of the associated processing may be performed prior to, concurrently with, and/or after the read operation). Furthermore, the term “identify” is used generically to describe any manner or mechanism for directly or indirectly ascertaining something, which may include, but is not limited to receiving, retrieving from memory, determining, defining, calculating, generating, etc.

Moreover, the terms “network” and “communications mechanism” are used generically herein to describe one or more networks, communications media or communications systems, including, but not limited to the Internet, private or public telephone, cellular, wireless, satellite, cable, local area, metropolitan area and/or wide area networks, a cable, electrical connection, bus, etc., and internal communications mechanisms such as message passing, interprocess communications, shared memory, etc. The term “message” is used generically herein to describe a piece of information which may or may not be, but is typically communicated via one or more communication mechanisms of any type.

The term “storage mechanism” includes any type of memory, storage device or other mechanism for maintaining instructions or data in any format. “Computer-readable medium” is an extensible term including any memory, storage device, storage mechanism, and other storage mechanisms that tangibly embodies computer-readable instructions and/or memory. The term “memory” includes any random access memory (RAM), read only memory (ROM), flash memory, integrated circuits, and/or other memory components or elements. The term “storage device” includes any solid state storage media, disk drives, diskettes, networked services, tape drives, and other storage devices. Memories and storage devices may store computer-executable instructions to be executed by a processing element and/or control logic, and data which is manipulated by a processing element and/or control logic. The term “data structure” is an extensible term referring to any data element, variable, data structure, database, and/or one or more organizational schemes that can be applied to data to facilitate interpreting the data or performing operations on it, such as, but not limited to memory locations or devices, sets, queues, trees, heaps, lists, linked lists, arrays, tables, pointers, etc. A data structure is typically maintained in a storage mechanism. The terms “pointer” and “link” are used generically herein to identify some mechanism for referencing or identifying another element, component, or other entity, and these may include, but are not limited to a reference to a memory or other storage mechanism or location therein, an index in a data structure, a value, etc.

The term “one embodiment” is used herein to reference a particular embodiment, wherein each reference to “one embodiment” may refer to a different embodiment, and the use of the term repeatedly herein in describing associated features, elements and/or limitations does not establish a cumulative set of associated features, elements and/or limitations that each and every embodiment must include, although an embodiment typically may include all these features, elements and/or limitations. In addition, the phrase “means for xxx” typically includes computer-readable medium containing computer-executable instructions for performing xxx.

In addition, the terms “first,” “second,” etc. are typically used herein to denote different units (e.g., a first element, a second element). The use of these terms herein does not necessarily connote an ordering such as one unit or event occurring or coming before another, but rather provides a mechanism to distinguish between particular units. Additionally, the use of a singular tense of a noun is non-limiting, with its use typically including one or more of the particular thing rather than just one (e.g., the use of the word “memory” typically refers to one or more memories without having to specify “memory or memories,” or “one or more memories” or “at least one memory”, etc.). Moreover, the phrases “based on x” and “in response to x” are used to indicate a minimum set of items x from which something is derived or caused, wherein “x” is extensible and does not necessarily describe a complete list of items on which the operation is performed, etc. Additionally, the phrase “coupled to” is used to indicate some level of direct or indirect connection between two elements or devices, with the coupling device or devices modifying or not modifying the coupled signal or communicated information. The term “subset” is used to indicate a group of all or less than all of the elements of a set. The term “subtree” is used to indicate all or less than all of a tree. Moreover, the term “or” is used herein to identify a selection of one or more, including all, of the conjunctive items.

Disclosed are, inter alia, methods, apparatus, data structures, computer-readable media, mechanisms, and means for identifying entries in a content-addressable memory that differ from a lookup word in less than a predetermined number of bit positions. Note, the embodiments disclosed herein can be used to provide a reliable lookup mechanism allowing for a small number of transient bit errors, wherein each of the original stored associative memory entries differ in more bit positions than the maximum number of bit position misses that which will still allow the associative memory entry to be identified as matching the lookup word, such as that disclosed in Shoham et al., “Error Protection for Lookup Operations in Content-Addressable Memory Entries,” U.S. patent application Ser. No. 10/879,254, filed Jun. 29, 2004, and hereby incorporated by reference in its entirety. Of course, embodiments of the invention disclosed herein can also be used in other applications.

One embodiment for identifying matching values differing in less than a predetermined number of bit positions includes: a content-addressable memory entry including n content-addressable memory entry cells, a match line, and a comparator. Each particular content-addressable memory entry cell is configured to identify whether or not a bit value of a lookup word in a corresponding bit position in the lookup word matches the particular content-addressable memory entry cell's predetermined bit value. The match line is configured to be precharged during a precharge phase prior to a matching phase of a lookup operation on the lookup word, and then to provide a match reference voltage responsive to the identifications of not matching by the content-addressable memory cells, such that the match reference voltage falls at rate corresponding to the number of content-addressable memory entry cells identifying it is not matched. The comparator is configured to identify whether or not the match reference voltage is above or below a predetermined voltage level selected to reflect a matching of the lookup word by at least m of the content-addressable memory cells at a predetermined time of the lookup operation, wherein the match reference voltage is selected such that it will be above the predetermined voltage level at the predetermined time if at least m of the content-addressable memory cells did not identify themselves as not matching and the match reference voltage will be below the predetermined voltage level at the predetermined time if at least n−m+1 of the content-addressable memory cells did identify themselves as not matching; wherein n>m>0. In one embodiment, the predetermined voltage level is static for the content-addressable memory entry (e.g., manufactured at a voltage level, a constant voltage level received from an external source, etc.). In one embodiment, the predetermined voltage level is dynamic to adjust for different values of m, such as, but not limited to a received voltage level of possibly different voltage levels, a selectable voltage level from different voltages levels, etc. One embodiment includes a data latching mechanism coupled to the comparator for latching the identification of whether or not the match reference voltage is above or below a predetermined voltage level at the predetermined time.

One embodiment for identifying matching values differing in less than a predetermined number of bit positions includes multiple content-addressable memory entries configured to identify whether or not they match a lookup word and a priority encoder configured to identify a highest priority matching entry from the entries indicating that it matched, if any. Each of the content-addressable memory entries includes: n content-addressable memory entry cells, a match line, and a comparator. Each particular content-addressable memory entry cell is configured to identify whether or not a bit value of a lookup word in a corresponding bit position in the lookup word matches the particular content-addressable memory entry cell's predetermined bit value. The match line is configured to be precharged during a precharge phase prior to a matching phase of a lookup operation on the lookup word, and then to provide a match reference voltage responsive to the identifications of not matching by the content-addressable memory cells, such that the match reference voltage falls at rate corresponding to the number of content-addressable memory entry cells identifying it is not matched. The comparator is configured to identify whether or not the match reference voltage is above or below a predetermined voltage level selected to reflect a matching of the lookup word by at least m content-addressable memory cells at a predetermined time of the lookup operation, wherein the match reference voltage is selected such that it will be above the predetermined voltage level at the predetermined time if at least m of the content-addressable memory cells did not identify themselves as not matching and the match reference voltage will be below the predetermined voltage level at the predetermined time if at least n−m+1 of the content-addressable memory cells did identify themselves as not matching; wherein n>m>0. In one embodiment, each of the content addressable memory entries includes a predetermined voltage level mechanism for generating a static voltage level for the predetermined voltage level. In one embodiment, each of the content addressable memory entries includes a predetermined voltage level mechanism for dynamically defining the predetermined voltage level corresponding to m. In one embodiment, each of the content-addressable memory entries includes a data latching mechanism coupled to the comparator for latching the identification of whether or not the match reference voltage is above or below a predetermined voltage level at the predetermined time.

One embodiment for identifying matching values differing in less than a predetermined number of bit positions, includes: a content-addressable memory entry including n content-addressable memory entry cells, each particular content-addressable memory entry cell being configured to identify whether or not a corresponding bit value of a lookup word in matches the particular content-addressable memory entry cell's predetermined bit value; means for providing a match reference voltage reflective of the number of the identifications by the content-addressable memory cells of whether or not they matched, with the match reference voltage varying during a matching phase of a lookup operation on the lookup word at a rate said reflective of the number of the identifications by the content-addressable memory cells of whether or not they matched; and means for identifying whether or not the match reference voltage reflects a match of at least m said content-addressable memory cells and not a match of less than m said content-addressable memory cells at a predetermined time during the matching phase; wherein n>m>0. One embodiment includes means for providing a predetermined voltage level for comparing with the match reference voltage level. In one embodiment, means for providing the predetermined voltage level includes circuitry for providing a static voltage level corresponding to m. In one embodiment, means for providing the predetermined voltage level includes circuitry for receiving a voltage level from an external source, with this voltage level being used directly or after adjustment as the predetermined voltage level. In one embodiment, means for providing the predetermined voltage level includes circuitry for selecting among different predetermined voltage levels corresponding to different values of m. In one embodiment, means for providing the predetermined voltage level includes circuitry for producing or generating different predetermined voltage levels corresponding to different values of m. One embodiment includes a data latching mechanism coupled to the comparator for latching said identification of whether or not the match reference voltage is above or below a predetermined voltage level at the predetermined time.

One embodiment for identifying matching values differing in less than a predetermined number of bit positions: precharges a match line coupled to each of n content-addressable memory cells of a content-addressable memory entry in preparation for performing a lookup operation based on a lookup word; determines whether or not each of the content-addressable memory cells matches a bit value at a corresponding bit position in the lookup word, with each of said content-addressable memory cells that determined that it does not match providing a discharge path for the precharged match line causing a match reference voltage of the match line to vary during the lookup operation at a rate reflective of the number of content-addressable memory cells that determined that it does not match; and identifies whether or not the match reference voltage reflects a match of at least m said content-addressable memory cells and not a match of less than m said content-addressable memory cells at a predetermined time during the lookup operation; wherein n>m>0. One embodiment, in identifying whether or not the match reference voltage reflects a match of at least m said content-addressable memory cells and not a match of less than m said content-addressable memory cells at the predetermined time during the lookup operation, includes comparing the match reference voltage to a predetermined voltage level and identifying a match if the match reference voltage is greater than the predetermined voltage level and identifying no match if the match reference voltage is less than the predetermined voltage level. One embodiment includes generating a predetermined voltage level based on the value of m; and wherein identifying whether or not the match reference voltage reflects a match includes comparing the match reference voltage with the predetermined voltage level. One embodiment latches the identification of whether or not the match reference voltage reflects a match of at least m said content-addressable memory cells and not a match of less than m said content-addressable memory cells at the predetermined time during the lookup operation.

One embodiment for identifying matching values differing in less than a predetermined number of bit positions, includes: means for precharging a match line of each of n content-addressable memory cells of a content-addressable memory entry in preparation of performing a lookup operation based on a lookup word; means for determining whether or not each of said content-addressable memory cells matches a bit value at a corresponding bit position in the lookup word, with each of the content-addressable memory cells determining that it does not match providing a discharge path for the match line previously precharged for performing the lookup operation on the lookup word causing a match reference voltage of the match line to vary during the lookup operation at a rate reflective of the number of said content-addressable memory cells that determined that it does not match; and means for identifying whether or not the match reference voltage reflects a match of at least m said content-addressable memory cells and not a match of less than m said content-addressable memory cells at a predetermined time during the lookup operation; wherein n>m>0. One embodiment includes means for generating a predetermined voltage level based on the value of m; wherein said means for identifying whether or not the match reference voltage reflects a match includes means for comparing the match reference voltage with the predetermined voltage level.

In one embodiment, n−m+1 equals two. In one embodiment, n−m+1 is more than two. In one embodiment, the bit value consists of two possible states corresponding to a zero and a one (such as in a binary content-addressable memory). In one embodiment, the bit value consists of three possible states corresponding to a zero, a one, and don't care (such as in a ternary content-addressable memory).

One embodiment is configured to identify matching values differing in less than a predetermined number of ternary content-addressable memory (TCAM) cells. A TCAM entry includes t TCAM entry cells, with each particular TCAM entry cell of said TCAM entry cells being configured to identify whether or not a bit value of a lookup word in a bit position in the lookup word corresponding to said particular TCAM cell matches the particular TCAM entry cell's predetermined value of a set of values including: one, zero and don't care. A match line, coupled to each of said TCAM entry cells, is configured to provide a match reference voltage after the match line is precharged during a precharge phase prior to a matching phase of a lookup operation on the lookup word. The match line is responsive to said identifications of not matching by said TCAM cells, such that the match reference voltage falls at rate corresponding to the number of active discharge paths in the TCAM entry cells that do not match for the lookup word. A comparator, coupled to the match line, is configured to identify whether or not the match reference voltage is above or below a predetermined voltage level selected to reflect a matching of the lookup word by at least t-u said TCAM cells at a predetermined time of the lookup operation. The predetermined voltage level is selected such that said match reference voltage will be above the predetermined voltage level at the predetermined time if u or less of said TCAM cells said identified as not matching and the match reference voltage will be below the predetermined voltage level at the predetermined time if at least u+1 said TCAM cells said identified as not matching; wherein t>u>0.

One embodiment performs error protected lookup operations on a data value in a ternary content-addressable memory (TCAM) entry. The TCAM entry includes t TCAM cells; the data value includes a plurality of data bit values from a set of data bit values including one, zero, and don't care; and at least 2u+1 copies of the data value are stored among the t TCAM cells. One embodiment includes a match line coupled to each of said t TCAM cells; means for precharging the match line coupled to the match line; means for determining whether or not each of said TCAM cells matches a bit value at a corresponding bit position in a lookup word, with each of said TCAM cells said determining that it does not match providing a discharge path for the previously precharged match line causing a match reference voltage of the match line to vary during the lookup operation at a rate reflective of the number of said TCAM cells that said determined that it does not match; and means for determining whether or not the match reference voltage reflects not matching u or less than u said TCAM cells at a predetermined time during the lookup operation thus determining whether or not the lookup word matched the TCAM entry; wherein t>u>0. In one embodiment, said means for determining whether or not the match reference voltage reflects not matching less than u said TCAM cells at a predetermined time during the lookup operation includes means for comparing the match reference voltage to a predetermined voltage level and identifying a hit if the match reference voltage is greater than the predetermined voltage level and identifying a miss if the match reference voltage is less than the predetermined voltage level.

One embodiment includes a data latching mechanism coupled to the comparator for latching said identification of whether or not the match reference voltage is above or below a predetermined voltage level at the predetermined time. One embodiment includes a predetermined voltage level mechanism for generating a static voltage level for the predetermined voltage level. One embodiment includes a predetermined voltage level mechanism for dynamically defining the predetermined voltage level corresponding to u. In one embodiment, u is at least two. In one embodiment, at least 2u+1 copies of a data word to be protected are stored among said t TCAM entry cells. One embodiment includes a storage word replicater such that at least 2u+1 copies of a received data word to be protected are stored among said t TCAM entry cells. In one embodiment, the lookup word includes 2u+1 copies of a portion of the lookup word. One embodiment provides 2u+1 copies of a portion of the lookup word to the TCAM entry in response to receiving the portion of the lookup word.

One embodiment performs error protected lookup operations on a data value in a ternary content-addressable memory (TCAM) entry. The TCAM entry includes t TCAM cells, and the data value includes a plurality of data bit values from a set of data bit values including one, zero, and don't care. At least 2u+1 copies of the data value are stored among said t TCAM cells, t>u>0. A match line coupled to each of said t TCAM cells is precharged in preparation for performing a lookup operation based on a lookup word. The lookup word is applied to said t TCAM cells. A determination is made as to whether or not each of said TCAM cells matches a bit value at a corresponding bit position in the lookup word, with each of said TCAM cells said determining that it does not match providing a discharge path for the match line previously said precharged match line causing a match reference voltage of the match line to vary during the lookup operation at a rate reflective of the number of said TCAM cells that said determined that it does not match. A hit/miss indication of whether or not the match reference voltage reflects not matching up to u said TCAM cells at a predetermined time during the lookup operation is provided to indicate whether or not the lookup word matched the TCAM entry.

In one embodiment, u is at least two. In one embodiment, the lookup word includes 2u+1 copies of a portion of the lookup word. In one embodiment, the match reference voltage is compared to a predetermined voltage level and a hit is identified if the match reference voltage is greater than the predetermined voltage level and a miss is identified if the match reference voltage is less than the predetermined voltage level.

One embodiment includes an associative memory cell configured to selectively produce a matching result of two binary content-addressable memory (CAM) bits or of one ternary content-addressable memory (TCAM) bit. In one embodiment, the associative memory cell comprises: two CAM bit cells configured to identify whether or not their associated programmed value matches a respective input value; match detection logic, coupled to said two CAM bit cells, configured to generate the matching result based on said matching of said two CAM bit cells; and selective circuitry, coupled to the two CAM bit cells, configured to selectively provide input search values as search values to said two CAM bit cells such that in a CAM bit mode of the selective circuitry, the two CAM bit cells provide two CAM bit results, and when in a TCAM bit mode, the two CAM bit cells provide a single TCAM bit result.

In one embodiment, said match detection logic includes: a match line; and a match detector coupled to the match line; wherein each of the two CAM bit cells are coupled to the match line and are configured to pull down the match line when their associated programmed value does not match its said respective search value; and the match detector is configured to said generate the matching result based on whether or not the associative memory cell identifies a match of said TCAM bit or both of said CAM bits based on a voltage level of the match line. In one embodiment, in a TCAM bit mode, said program values associated with the two CAM bit cells are a same value when the TCAM bit represents a don't care value. In one embodiment, wherein in a TCAM bit mode, said program values associated with the two CAM bit cells are different values when the TCAM bit represents a one or a zero value. In one embodiment, the associative memory is disposed on a single chip.

One embodiment includes an apparatus configured to match input values. In one embodiment, the apparatus comprises: a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values; circuitry configured to selectively apply the plurality of search values to the plurality of associative memory bit cells in a manner such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory (CAM) cells or (b) m ternary content-addressable memory (TCAM) cells, wherein n>m>0; wherein said selective application of said search values provides different values to the plurality of associative memory bit cells when in the n CAM mode than in the m TCAM mode; and a match detector coupled to the plurality of associative memory bit cells, the match detector configured to identify whether or not at least one of the plurality of associative memory bit cells does not match their respective said search value in order to produce the n CAM cell or m TCAM cell matching result.

In one embodiment, n is two and m is one. In one embodiment, in a TCAM bit mode, said program values associated with the two CAM bit cells are a same value when the TCAM bit represents a don't care value. In one embodiment, in a TCAM bit mode, said program values associated with the two CAM bit cells are different values when the TCAM bit represents a one or a zero value. In one embodiment, the associative memory is disposed on a single chip.

One embodiment includes an associative memory configured to match input values. In one embodiment, the associative memory comprises: a plurality of rows of associative memory cells and circuitry configured to selectively provide input search values to the associative memory cells. In one embodiment, each particular row of the plurality of rows includes: one or more associative memory mode identification cells, each configured to store a value representing that said particular row is either in a binary content-addressable memory (CAM) bit mode or a ternary content-addressable memory (TCAM) bit mode, and to determine whether or not its said stored value matches a search operation mode value identifying whether a current search is to use either (a) the rows programmed in CAM bit mode or (b) the rows programmed in TCAM bit mode; a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values; and a match detector configured to, in response to identifying that said one or more associative memory mode identification cells in its said particular row or any of the plurality of associative memory bit cells in its said particular row identifies itself as currently not matching, identify said particular row as not matching, else identify said particular row as matching. In one embodiment, the circuitry is configured to selectively provide, in response to the search operation mode value, the plurality of search values to the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells in a manner such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory cells in response to a value of the search operation mode value being a first value, or (b) m ternary content-addressable memory (TCAM) cells response to a value of the search operation mode value being a different value than the first value, wherein n>m>0; wherein said selective application of said search values provides different values to the plurality of associative memory cells when in the n CAM mode than in the m TCAM mode.

In one embodiment, n is two and m is one, and each of the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells includes two CAM bit cells. In one embodiment, the associative memory is disposed on a single chip.

One embodiment includes an associative memory configured to match input values. In one embodiment, the associative memory comprises: a plurality of rows of associative memory cells and circuitry configured to selectively provide input search values to the associative memory cells. In one embodiment, each particular row of the plurality of rows includes: a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values; one or more associative memory mode identification cells each configured to store a value representing that an associated group of said associative memory bit cells in the particular row is either in a binary content-addressable memory (CAM) bit mode or a ternary content-addressable memory (TCAM) bit mode; and a match detector configured to, in response to identifying that any of the plurality of associative memory bit cells in its said particular row identifies itself as currently not matching, identify said particular row as not matching, else identify said particular row as matching. In one embodiment, the circuitry is configured to selectively provide the plurality of search values to the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells in a manner according to said associative memory mode identification cells for said groups of associative memory bit cells such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory cells or (b) m ternary content-addressable memory (TCAM) cells, wherein n>m>0. In one embodiment, n is two and m is one, and each of the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells includes two CAM bit cells. In one embodiment, the associative memory is disposed on a single chip.

One embodiment includes an associative memory configured to match input values with a predetermined error budget of e bit errors. In one embodiment, the associative memory comprises: a plurality of rows of associative memory cells and circuitry configured to selectively provide input search values to the associative memory cells. In one embodiment, each particular row of the plurality of rows includes: at least 2e+1 associative memory mode identification cells, each configured to store a value representing that said particular row is either in a binary content-addressable memory (CAM) bit mode or a ternary content-addressable memory (TCAM) bit mode, and to determine whether or not its said stored value matches a search operation mode value identifying whether a current search is to use either (a) the rows programmed in CAM bit mode or (b) the rows programmed in TCAM bit mode; a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values; and a match detector configured to, in response to identifying that more than e of said one or more associative memory mode identification cells in its said particular row or more than e of the plurality of associative memory bit cells in its said particular row identify themselves as currently not matching, identify said particular row as not matching, else identify said particular row as matching. In one embodiment, the circuitry is configured to selectively provide, in response to the search operation mode value, the plurality of search values to the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells in a manner such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory cells in response to a value of the search operation mode value being a first value, or (b) m ternary content-addressable memory (TCAM) cells response to a value of the search operation mode value being a different value than the first value, wherein n>m>0; wherein said selective application of said search values provides different values to the plurality of associative memory cells when in the n CAM mode than in the m TCAM mode.

In one embodiment, n is two and m is one, and each of the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells includes two CAM bit cells. In one embodiment, e is one. In one embodiment, the associative memory is disposed on a single chip.

One embodiment performs operations for identifying matches in an associative memory cell configured to selectively produce a matching result of two binary content-addressable memory (CAM) bits or of one ternary content-addressable memory (TCAM) bit, the associative memory cell comprising two CAM bit cells configured to identify whether or not their associated programmed value matches a respective search value; and match detection logic, coupled to said two CAM bit cells, configured to generate the matching result based on said matching of said two CAM bit cells against said search values. In one embodiment, these operations include: programming each of the two CAM bit cells with values to be matched against; and selectively providing search values to said two CAM bit cells based on one or more input search values and whether the associative memory cell is in a CAM or TCAM bit mode, such that in a CAM bit mode the two CAM bit cells produce two CAM bit results, and when in a TCAM bit mode, the two CAM bit cells produce a single TCAM bit result; wherein said selectively providing search values would provide different search values to said two CAM bit cells if the associative memory was in a different CAM or TCAM bit mode for the same said one or more input search values.

In one embodiment, said programming of each of the two CAM bit cells includes programming a same value to each of each of the two CAM bit cells when in the TCAM bit mode and the TCAM bit represents a don't care value. In one embodiment, said programming of each of the two CAM bit cells includes programming different values to said two CAM bit cells when in the TCAM bit mode and the TCAM bit represents a zero or a one. One embodiment includes means for performing said operations.

FIG. 2 illustrates an observed varying of the match reference voltage 200 of a precharged match line of a binary or ternary content-addressable memory cell in response to the number of content-addressable memory bit cells not matching their respective bits of a lookup word. As shown, depending on the number of misses (i.e., the number of comparison circuits providing a discharge path in the content-addressable memory entry), the rate at which the match reference voltage 200 will drop varies.

FIGS. 3A-B illustrate a mechanism for determining a predetermined voltage level and a predetermined time corresponding to the maximum number of bit positions that can be identified as not matching while still considering that the binary or ternary content-addressable memory entry matched for the lookup word. FIG. 3A illustrates (by the voltage drop curves 304-307) the varying of the match reference voltage 300 of a precharged match line of a content-addressable memory cell in response to the number of content-addressable memory bit cells not matching their respective bits of a lookup word. Each of the rate of voltage drop curves 304-307 correspond to varying numbers of content-addressable memory bit cells not matching their respective bits of a lookup word, with regions (311, 312, 313) being defined by these voltage drop curves 304-307. Of course, the actual discharge rate of a match line is dependent on the implementation of the embodiment, and therefore, the selection of the predetermined voltage level and predetermined time for sampling the match reference voltage level on the match line will need to be selected accordingly. A convenient mechanism for selecting these values is to simulate the circuitry of the implementation using a design tool (e.g., Spice) to plot the change in the match reference voltage on the match line for varying numbers of discharge paths produced in result of non-matching bit positions in a content-addressable memory entry.

Of particular interest in FIG. 3A is voltage drop curve 305, which corresponds to n−m+1 of the content-addressable memory entry bit cells identifying as not matching their corresponding bit of the lookup word, and drop curve 306, which corresponds to n−m of the content-addressable memory entry bit cells identifying as not matching their corresponding bit of the lookup word. Voltage drop curves 305 and 306 define region 312, wherein the intersection of the predetermined voltage level and predetermined time of the lookup operation should lie for comparing the match reference voltage of the match line to the predetermined voltage level for determining whether or not the content-addressable memory entry should be identified as matched for the lookup word.

This is further illustrated in FIG. 3B, which illustrates two possible intersection points 345 and 355, which could be used in an embodiment, as could other points within region 312. Typically, the intersection point is chosen for speed reasons (e.g., to have a shorter predetermined time for sampling the match line reference voltage) as well as to provide a margin for error within the tolerances of the actual devices used to implement the embodiment. As shown, intersection point 345 corresponds to predetermined voltage level-A 340 and predetermined time-A 341, and intersection point 355 corresponds to predetermined voltage level-B 350 and predetermined time-B 351.

FIG. 4A is a block diagram of a binary or ternary content-addressable memory entry 400 of a content-addressable memory according to one embodiment. As shown, content-addressable memory entry 400 includes multiple CAM/TCAM cells including, but not limited to, CAM/TCAM cells 401A-B. Each cell 401A-B typically includes a storage for storing a value (e.g., one, zero, don't care) to be matched against (although in one embodiment, these values are stored elsewhere). In order to determine if content-addressable memory entry 400 matches a lookup word, match line 432 is precharged to Vdd based on precharging signal 420 during a precharging phase of the lookup operation. The encode bit values (e.g., typically the value and its complement for a CAM or the encoded value or complement thereof for a TCAM) of the lookup word are applied on bit lines 411A-B and 412A-B, which are compared by comparison circuits 430 during a matching phase of the lookup operation to the bit values of bit cells 401A-B. For a ternary content-addressable memory, the comparison operation is effectively disabled if the bit cell is programmed with a don't care—as it matches any value of the corresponding bit of the lookup word. If any of the entry bit cells 401A-B determine that they do not match, then match line 432 is discharged by the corresponding comparison circuit(s) 430 at a rate reflective of the number of entry bit cells 401A-B identifying that their value does not match their respective bit of the lookup word (such as, but not limited to that illustrated in FIGS. 2, 3A, and 3B), with each of these non-matching bit cells providing a discharge path for precharged match line 432.

Comparator 435 compares the match reference voltage on match line 432 to predetermined voltage level 433 at the predetermined sample time 438 (with these predetermined values selected to provide proper results, such as selected in a manner disclosed herein). In one embodiment, predetermined voltage level mechanism 434 generates a static predetermined voltage level 433. In one embodiment, a voltage level is received from an external source, and predetermined voltage level mechanism 434 relays, possibly after adjusting the received voltage level, to comparator 435 as predetermined voltage level 433. In one embodiment, predetermined voltage level 433 is selected from among different voltage levels corresponding to different values of m, and predetermined voltage level mechanism 434 includes mechanisms for selecting among different voltage levels and/or for producing different voltage levels. For example, in one embodiment, a value reflective of m is received by predetermined voltage level mechanism 434, which corresponding defines (e.g., selects, produces, generates, etc.) predetermined voltage level 433 provided to comparator 435.

Comparators are well-known in the art, with a good reference being Phillip E. Allen and Douglas R. Holberg, CMOS Analog Circuit Design, Oxford University Press (Second Ed.), 2002, Chapter 8. In one embodiment comparator 435 is responsive to predetermined sample time signal 438 to produce hit/miss indication 436 based on the comparison at the predetermined time (as indicated by signal 438). In one embodiment, comparator 435 continuously compares the match reference voltage on match line 432 to predetermined voltage level 433 to continuously generate hit/miss indication 436, while latching mechanism 437 is responsive to latch the value of hit miss indication 436 at predetermined sample time 438 (e.g., exactly at the time or shortly thereafter to allow for settling of the results of the comparison and/or a propagation delay), and to generate hit/miss indication 439 reflective of the comparison between the match reference voltage on match line 432 and predetermined voltage level 433 at the predetermined time (as indicated by signal 438).

FIG. 4B illustrates a process used in one embodiment. Processing begins with process block 460, and proceeds to process block 462, wherein the match line of each of the binary or ternary content-addressable memory entries is precharged. In process block 464, each bit cell of each content-addressable memory entry (or at least the enabled bit cells not having a don't care bit value) identifies whether or not it matches its corresponding bit in a lookup word, and if it does not match, it provides a discharge path for the match reference voltage of the precharged match line. As determined in process block 465, when the predetermined time arrives, then in process block 466, a comparison is made between the current value of the match reference voltage of the match line and the predetermined voltage level to identify whether or not the entry should be identified as being matched. In one embodiment, the match reference voltage is a static voltage level. In one embodiment in process block 466, the predetermined voltage level is defined (e.g., selected, produced, generated, etc.) corresponding to the desired value of m. As determined in process block 470, if any of the content-addressable entries are identified as being matched, then in process block 474, the highest priority one of the matching content-addressable entries is identified; otherwise, a no-match indication is identified in process block 472. Processing is complete as indicated by process block 475.

FIG. 5 is a block diagram of a binary or ternary content-addressable memory 500 used in one embodiment. As shown, content-addressable memory 500 includes multiple content-addressable memory entries 510, each of which is configured as disclosed herein (of course, or an equivalent thereof) to identify whether or not the entry should be considered to have matched lookup word 521 even if less than all bit positions match. Control logic 520 provides control signal 523 (e.g., precharge, predetermined sample time, read/write line, etc.) to content-addressable memory entries 510 and to priority encoder 530 to control the timing and lookup operations. Priority encoder 530 typically identifies the highest priority one of content-addressable memory entries 510 which are identified as matching lookup word 521 or a no-hit indication for the current lookup operation to produce indication 531.

FIG. 6 is a flow diagram of a process used in one embodiment for programming TCAM entries. In one embodiment, each TCAM entry includes t TCAM cells. Conventional TCAMs store a data value, which includes a plurality of data bit values from a set of data bit values including one, zero, and don't care. By storing only the data value, no error protection is available. In other words, these conventional TCAMs are designed with a value of u of zero (i.e., all cells of an entry must match in order for a TCAM entry to be considered as matching).

In contrast, one embodiment, not only stores the original data value, but stores an additional 2u copies of the data value to provide a margin wherein up to u TCAM entry cells can have errors, yet only true hits with no false hits or false misses will be detected. Note, the phrase “copies of the data value” refers to values being exactly the same, or some deterministic value derived there from (e.g., a one being mapped to a zero, and vice versa).

Therefore, in one embodiment, a TCAM entry is programmed with exactly, or at least 2u+1 copies of the data value stored among said t TCAM cells, wherein u refers to the maximum number of non-matching TCAM cells that will be allowed while still considering the TCAM entry to be matching. The number of protection bits required for protecting TCAM cells is typically larger than that required for binary CAM cells because of the implications of the don't care value, which can be programmed in or result from a bit error in the TCAM cell. Also, the don't care value allows the matching of different lookup words by multiple different TCAM entries. By storing an additional 2u copies of each bit of a data value in a TCAM entry, error free protection of up to u bit errors is provide. Therefore, in order to provide one bit of error protection (i.e., u=1) for a data value of length twenty-four, a TCAM entry will require at least three (2u+1) times twenty-four or seventy-two TCAM entry cells. Note, these seventy-two TCAM entry cells can be in any order, as long as the 2u+1 (i.e., seventy-two bits in this example) bits of the lookup word are applied in the same ordering. One embodiment automatically generates the 2u copies of the received lookup word and correspondingly applies them to the TCAM entry cells. The (2u+1-sized) lookup word is applied to the t TCAM cells. A determination is made as to whether or not at least all but u of said TCAM cells match their respective bit position in the lookup word. A hit/miss indication of whether or not at least all but u of said TCAM cells match their respective bit position in the lookup word is generated. In one embodiment, t>u>0. In one embodiment, this determination is performed using discrete logic. In one embodiment, this determination is made based on a voltage level of one or more match lines.

FIG. 6 is a flow diagram of a process used in one embodiment for programming TCAM entries. Processing beings with process block 600. As determined in process block 602, while there are more data values to store: the next data value is identified (604), and the 2u+1 copies of the data value are stored in the TCAM entry among its t TCAM cells (606). When all data values have been programmed, processing is complete as indicated by process block 609.

FIG. 7A illustrates an associative memory bit cell 700 according to one embodiment. As shown, associative memory bit cell 700 can perform lookup operations acting as two CAM bits or a single TCAM bit, based on the search values 713, 714 provided to CAM bit cells 701, 702, (and of course, the programming of the values associated with CAM bit cells 701, 702). In one embodiment, search mode signal 710 is used to identify whether the search being currently performed is on the associative memory bit cells configured as n CAM bit cells or m TCAM bit cells, such that circuitry can provide the appropriate values to CAM bit cells 701, 702 based on search values 713, 714. Match line 713 is initially precharged (713), and then if either CAM bit cells 701, 702 do not match, then they pull match line 713 low. As shown, a match line sense amplifier 718 can be used to generate the resulting hit or miss indication 719. In one embodiment, a voltage comparator is used to generate the resulting hit or miss indication 719.

FIG. 7B illustrates an associative memory bit cell 720 according to one embodiment. As shown, associative memory bit cell 720 can perform lookup operations acting as two CAM bits or a single TCAM bit, based on the input search values 727, 728. In one embodiment, search mode signal 723 is used to identify whether the search being currently performed is on the associative memory bit cells configured as n CAM bit cells or m TCAM bit cells, such that circuitry 724 can provide the appropriate search values 726 to CAM bit cells 721, 722 based on input search values 727, 728. If either CAM bit cells 721, 722 do not match, then they pull match line 725 low.

In one embodiment, when the search mode (723) indicates a CAM lookup operation mode, then circuitry 724 provides the value of input search value 727 and its complement to CAM bit cell 721 and provides the value of input search value 728 and its complement to CAM bit cell 722. When search mode (723) indicates a TCAM lookup operation mode, then circuitry 724 provides the value of input search value 727 and a low voltage level to CAM bit cell 721 and provides the value of input search value 728 and a low voltage level to CAM bit cell 722, such that the combination of CAM bit cells 721, 722 operate as a single TCAM bit cell.

FIG. 7C illustrates that one embodiment shares a single set of circuitry (730) configured to provide search values (731) to multiple rows (732-733) of associative memory bit cells, with each of the associative memory cells being configured to operate in response to corresponding search values of search values 731.

The programming and operation of one embodiment is further illustrated in relation to FIG. 7D, which describes (750) one way of programming the CAM bit cells and for providing the search values to the CAM bit cells, wherein n=2 and m=1 (i.e., the associative memory bit cells operate as either two CAM bits or one TCAM bit).

When acting as two CAM bits, each CAM bit cell 721, 722 (FIG. 7B) is programmed with their desired values, with these standard values being provided to each CAM bit cell 721, 722 as their search values. Thus, in CAM operating mode, a CAM cell 721, 722 programmed with a ‘1’ will match an input ‘1’ value and will not match an input ‘0’ value; and a CAM cell 721, 722 programmed with a ‘0’ will match an input ‘0’ value and will not match an input ‘1’ value.

When acting as a single TCAM bit, each CAM bit cell 721, 722 (FIG. 7B) is programmed (750) as shown, with a ‘0’ being represented as ‘01’, a ‘1’ being represented as ‘10’, and a don't care case being represented as ‘00’. Therefore, both CAM bit cells 721, 722 will be prevented from providing a discharge path to pull down match line 725. If a search value for the TCAM bit is ‘0’, then the search values of ‘01’ are provided to CAM bit cells 721, 722; while if a search value for the TCAM bit is ‘1’, then the search values of ‘10’ are provided to CAM bit cells 721, 722. Therefore, if there is not a match, either CAM bit cell 721 or 722 will provide discharge path to pull down match line 725; while if there is a match, no discharge path will be provided to pull down match line 725.

Of course, as contemplated by the invention and would be understood by one skilled in the art, the mapping of values to stored bits and lookup values could be reversed with the same results achieved, and if different transistor types (e.g., NMOS) were used, the mapping of these values would need to be adjusted to match the implementation technology. Also, of course, enable logic is provided in one embodiment to determine whether or not an associative memory bit cell is to participate in a current search or whether it is programmed, etc.

FIG. 7E illustrates a flow diagram of a process used in one embodiment for identifying whether or not an associative memory bit cell should be considered as matching. Processing begins with process block 760, and proceeds to process block 762, wherein each of the CAM bit cells of the associative memory bit cell is programmed based on whether it is to be used in CAM or TCAM mode. In process block 764, the input search values are identified. As determined in process block 765, if TCAM lookup mode is to be used, then in process block 766, search values, based on the input search values, are provided to each of the CAM bit cells according to the TCAM encoding scheme. Otherwise, in process block 768, search values, based on the input search values, are provided to each of the CAM bit cells according to the CAM encoding scheme. As determined in process block 770, if any CAM bit cell does not match, then in process block 772, the associative memory bit cell is identified as not matching; otherwise, in process block 774, the associative memory bit cell is identified as matching. Processing typically returns to process block 764 to process more input values.

FIG. 8A is a block diagram of a row 800 including multiple associative memory bit cells 802-803, as used in one embodiment. As illustrated, one embodiment connects multiple associative memory bit cells 802-803 to a common match line 815, such that if any does not identify a match, such as, but not limited to that described in relation to FIGS. 7A-D, then match line 815 is pulled low.

In addition, one embodiment adds to row 800 an associative memory identification cell (e.g., a standard CAM cell), which is used to store a value corresponding whether the row 800 including associative memory bit cells 802-803 is configured to act as CAM or TCAM bits. Therefore, if the value of associative memory identification cell 801 does not match the search mode value 811 (e.g., whether to perform a search on CAM configured or TCAM configured associative memory bit cells), then match line 815 will be pulled low (effectively causing row 800 not to be matched against). Otherwise, if memory identification cell 801 matches search mode value 811, then whether or not a match is determined by row 800 will be determined by the result of the matching performed by associative memory bit cells 802-803.

This operation is further described in relation to FIG. 8B, which illustrates a flow diagram of a process used in one embodiment for identifying whether or not a row of associative memory bit cells should be considered as matching. Processing begins with process block 820, and proceeds to process block 821, wherein each of the CAM bit cells of the associative memory bit cell and the associative memory identification cell is programmed based on whether the row is to be used in CAM or TCAM mode. In process block 824, the input search values, including a search mode value, are identified. As determined in process block 825, if TCAM lookup mode is to be used, then in process block 826, search values, based on the input search values, are provided to each of the CAM bit cells according to the TCAM encoding scheme. Otherwise, in process block 828, search values, based on the input search values, are provided to each of the CAM bit cells according to the CAM encoding scheme. As determined in process block 830, if any CAM bit cell does not match including the associative memory identification cell matching the search mode value, then in process block 832, the associative memory bit cell is identified as not matching; otherwise, in process block 834, the associative memory bit cell is identified as matching. Processing typically returns to process block 824 to process more input values.

The usefulness of such features illustrated in relation to FIGS. 8A-B are further illustrated in relation to FIGS. 8C-D. FIG. 8C illustrates an associative memory 860, wherein each row 871-873 can be programmed to act as either a row of CAM cells or TCAM cells, with associative memory identification cell programmed accordingly. Input values 861-864 are provided to the respective cells in each row 871-873, with their search results provided to sense amplifier and priority encoder (868), which generates the resultant lookup result 869. One embodiment of associative memory 860 includes one set of circuitry to provide the search values to each column of the associative memory cells based on the search mode and the input search values (rather than requiring each associative memory cell to include such circuitry, e.g., circuitry 724 of FIG. 7B).

Therefore, each row can independently be configured as n CAM bits or as m CAM bits, with n>m>0. Therefore, when n=2 and m=1, a row configured to act in CAM mode will allow for up to twice the number of CAM bits than the number of TCAM bits (if configured in TCAM mode). Other embodiments implemented with different values of n and m may have a different ratio of possible CAM cells to TCAM cells. For example, this may be of particular use in a packet switching device which could use a same dual use associative memory 860 to support IP routing (e.g., using TCAM mode) and bridge routing (e.g., using CAM mode), with allowing for a bridge address to be twice the length of a IP routing address.

FIG. 8D illustrates that one embodiment, rather than configuring an entire rows of associative memory bit cells together as either all being CAM or TCAM bit cells as illustrated in FIGS. 8A and 8C, provides groupings 881-882 of subsets of associative memory bit cells within a row (or column) of associative memory bit cells. In one embodiment, these groupings 881-882 include associative memory identification cells (in which case typically multiple mode signals are provided), while in one embodiment they do not. Such a configuration may be advantageous for certain applications in which it is desirous to exactly match one or more portions of a row while allowing TCAM matching for one or more portions of the row.

FIG. 9A illustrates another embodiment in which error protection functionality (as already described herein) can be provided to a dual use associative memory, with its associative memory cells being configurable to act as either n CAM cells or m TCAM cells. As shown, one embodiment includes error protected associative memory 900 that has its rows 911-913 individually configurable as either in CAM mode or in TCAM mode. However, instead of detecting a miss if a single occurrence of not matching is detected in a row, rather an error budget of e non-matches is allowed, and therefore, error protected associative memory use a voltage comparator or other logic to identify such matching within the error budget. Such programming and matching of the associative memory cells is discussed supra. However, associative memory identifications cells are used to identify the mode of each row (or grouping), then it is typically desirous to also provide error protection for the associative memory identifications cells in each row 911-913. Therefore, one embodiment includes 2e+1 associative memory identification cells in each row 911-913 in order to ensure that they are appropriately protected in either operating mode. Therefore, when e=1, then three associative memory identification cells are included in each row 911-913.

FIG. 9B is a block diagram of an embodiment including multiple associative memory bit cells. In one embodiment, each row has one dedicated associative memory identification cell (920) and multiple associative memory bit cells (921-922). In one embodiment, one or more of associative memory bit cells (921-922) are programmed to perform the function of an associative memory identification cell, while a different one or more of associative memory bit cells (921-922) are programmed to perform the function of an associative memory bit cells as described herein. This flexibility allows a single hardware implementation to operate with a programmable value of the error budget. For example, in one embodiment, at least 2e+1 associative memory identification cells are used in each row in order to ensure appropriate error protection for e bit errors whether operating in either CAM/TCAM operating mode. Therefore, when e=1, typically three associative memory identification cells are programmed for each row; when e=2, typically five associative memory identification cells are programmed for each row, etc. Of course, when e=0, the dedicated associative memory identification cell is typically used to identify whether the row is programmed as a CAM or TCAM row, such as, but not limited to that described in relation to FIGS. 8A-B.

In one embodiment, each row configured as a CAM row has at least 2e+1 associative memory identification cells programmed with the value of one, and each row configured as a TCAM row has at least 2e+1 associative memory identification cells programmed with the value of zero. When a CAM search is performed, each of the at least 2e+1 associative memory identification cells receive a search value of one; while when a TCAM search is performed, the dedicated associative memory identification cell receives a search value of zero and all other at least 2e associative memory identification cells receive a search value of one. Assuming there are no bit errors in the stored values in the programmed associative memory identification cells, then when a CAM search is performed: all associative memory identification cells programmed for CAM mode will match, therefore, creating no discharge paths; while all associative memory identification cells programmed for TCAM mode will not match guaranteeing at least 2e+1 discharge paths for the row, which can be identified as not matching by the row's voltage comparator. Also, when a TCAM search is performed: all associative memory identification cells programmed for TCAM mode will match (as the dedicated associative memory identification cell is zero and the associative memory bit cells acting as associative memory identification bit cells are programmed with zeros which will match any input value), therefore, creating no discharge paths, while all associative memory identification cells programmed for CAM mode will not match guaranteeing at least 2e+1 discharge paths for the row, which can be identified as not matching by the row's voltage comparator.

In view of the many possible embodiments to which the principles of our invention may be applied, it will be appreciated that the embodiments and aspects thereof described herein with respect to the drawings/figures are only illustrative and should not be taken as limiting the scope of the invention. For example and as would be apparent to one skilled in the art, many of the process block operations can be re-ordered to be performed before, after, or substantially concurrent with other operations. Also, many different forms of data structures could be used in various embodiments. The invention as described herein contemplates all such embodiments as may come within the scope of the following claims and equivalents thereof.

Claims

1. An associative memory cell configured to selectively produce a matching result of two binary content-addressable memory (CAM) bits or of one ternary content-addressable memory (TCAM) bit, the associative memory cell comprising: two CAM bit cells configured to identify whether or not their associated programmed value matches a respective input value;match detection logic, coupled to said two CAM bit cells, configured to generate the matching result based on said matching of said two CAM bit cells; andselective circuitry, coupled to the two CAM bit cells, configured to selectively provide input search values as search values to said two CAM bit cells such that in a CAM bit mode of the selective circuitry, the two CAM bit cells provide two CAM bit results, and when in a TCAM bit mode, the two CAM bit cells provide a single TCAM bit result.
2. The associative memory cell of claim 1, wherein said match detection logic includes: a match line; and a match detector coupled to the match line; wherein each of the two CAM bit cells are coupled to the match line and are configured to pull down the match line when their associated programmed value does not match its said respective search value; and the match detector is configured to said generate the matching result based on whether or not the associative memory cell identifies a match of said TCAM bit or both of said CAM bits based on a voltage level of the match line.
3. The associative memory cell of claim 1, wherein in a TCAM bit mode, said program values associated with the two CAM bit cells are a same value when the TCAM bit represents a don't care value.
4. The associative memory cell of claim 3, wherein in a TCAM bit mode, said program values associated with the two CAM bit cells are different values when the TCAM bit represents a one or a zero value.
5. The associative memory cell of claim 1, wherein in a TCAM bit mode, said program values associated with the two CAM bit cells are different values when the TCAM bit represents a one or a zero value.
6. The associative memory cell of claim 1, wherein the associative memory is disposed on a single chip.
7. An apparatus configured to match input values, the apparatus comprising: a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values;circuitry configured to selectively apply the plurality of search values to the plurality of associative memory bit cells in a manner such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory (CAM) cells or (b) m ternary content-addressable memory (TCAM) cells, wherein n>m>0; wherein said selective application of said search values provides different values to the plurality of associative memory bit cells when in the n CAM mode than in the m TCAM mode; anda match detector coupled to the plurality of associative memory bit cells, the match detector configured to identify whether or not at least one of the plurality of associative memory bit cells does not match their respective said search value in order to produce the n CAM cell or m TCAM cell matching result.
8. The apparatus of claim 7, wherein n is two and m is one.
9. The apparatus of claim 7, wherein in a TCAM bit mode, said program values associated with the two CAM bit cells are a same value when the TCAM bit represents a don't care value.
10. The apparatus of claim 9, wherein in a TCAM bit mode, said program values associated with the two CAM bit cells are different values when the TCAM bit represents a one or a zero value.
11. The apparatus of claim 10, wherein n is two and m is one.
12. The apparatus of claim 7, wherein in a TCAM bit mode, said program values associated with the two CAM bit cells are different values when the TCAM bit represents a one or a zero value.
13. The apparatus of claim 7, wherein the associative memory is disposed on a single chip.
14. An associative memory configured to match input values, the associative memory comprising: a plurality of rows of associative memory cells, wherein each particular row of the plurality of rows includes: one or more associative memory mode identification cells, each configured to store a value representing that said particular row is either in a binary content-addressable memory (CAM) bit mode or a ternary content-addressable memory (TCAM) bit mode, and to determine whether or not its said stored value matches a search operation mode value identifying whether a current search is to use either (a) the rows programmed in CAM bit mode or (b) the rows programmed in TCAM bit mode;a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values; anda match detector configured to, in response to identifying that said one or more associative memory mode identification cells in its said particular row or any of the plurality of associative memory bit cells in its said particular row identifies itself as currently not matching, identify said particular row as not matching, else identify said particular row as matching; andcircuitry configured to selectively provide, in response to the search operation mode value, the plurality of search values to the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells in a manner such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory cells in response to a value of the search operation mode value being a first value, or (b) m ternary content-addressable memory (TCAM) cells response to a value of the search operation mode value being a different value than the first value, wherein n>m>0; wherein said selective application of said search values provides different values to the plurality of associative memory cells when in the n CAM mode than in the m TCAM mode.
15. The associative memory of claim 14, wherein n is two and m is one, and each of the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells includes two CAM bit cells.
16. The associative memory of claim 14, wherein the associative memory is disposed on a single chip.
17. An associative memory configured to match input values, the associative memory comprising: a plurality of rows of associative memory cells, wherein each particular row of the plurality of rows includes: a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values;one or more associative memory mode identification cells each configured to store a value representing that an associated group of said associative memory bit cells in the particular row is either in a binary content-addressable memory (CAM) bit mode or a ternary content-addressable memory (TCAM) bit mode; anda match detector configured to, in response to identifying that any of the plurality of associative memory bit cells in its said particular row identifies itself as currently not matching, identify said particular row as not matching, else identify said particular row as matching; andcircuitry configured to selectively provide the plurality of search values to the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells in a manner according to said associative memory mode identification cells for said groups of associative memory bit cells such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory cells or (b) m ternary content-addressable memory (TCAM) cells, wherein n>m>0.
18. The associative memory of claim 17, wherein n is two and m is one, and each of the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells includes two CAM bit cells.
19. The associative memory of claim 17, wherein the associative memory is disposed on a single chip.
20. An associative memory configured to match input values with a predetermined error budget of e bit errors, the associative memory comprising: a plurality of rows of associative memory cells, wherein each particular row of the plurality of rows includes: at least 2e+1 associative memory mode identification cells, each configured to store a value representing that said particular row is either in a binary content-addressable memory (CAM) bit mode or a ternary content-addressable memory (TCAM) bit mode, and to determine whether or not its said stored value matches a search operation mode value identifying whether a current search is to use either (a) the rows programmed in CAM bit mode or (b) the rows programmed in TCAM bit mode;a plurality of associative memory bit cells configured to determine whether or not their associated value matches a respective search value of a plurality of search values; anda match detector configured to, in response to identifying that more than e of said one or more associative memory mode identification cells in its said particular row or more than e of the plurality of associative memory bit cells in its said particular row identify themselves as currently not matching, identify said particular row as not matching, else identify said particular row as matching; andcircuitry configured to selectively provide, in response to the search operation mode value, the plurality of search values to the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells in a manner such that the plurality of associative memory bit cells produce results as being either (a) n binary content-addressable memory cells in response to a value of the search operation mode value being a first value, or (b) m ternary content-addressable memory (TCAM) cells response to a value of the search operation mode value being a different value than the first value, wherein n>m>0; wherein said selective application of said search values provides different values to the plurality of associative memory cells when in the n CAM mode than in the m TCAM mode.
21. The associative memory of claim 20, wherein n is two and m is one, and each of the plurality of associative memory bit cells in each of the plurality of rows of associative memory cells includes two CAM bit cells.
22. The associative memory of claim 21, wherein e is one.
23. The associative memory of claim 20, wherein e is one.
24. The associative memory of claim 20, wherein the associative memory is disposed on a single chip.
25. A method for identifying matches in an associative memory cell configured to selectively produce a matching result of two binary content-addressable memory (CAM) bits or of one ternary content-addressable memory (TCAM) bit, the associative memory cell comprising two CAM bit cells configured to identify whether or not their associated programmed value matches a respective search value; and match detection logic, coupled to said two CAM bit cells, configured to generate the matching result based on said matching of said two CAM bit cells against said search values, the method comprising: programming each of the two CAM bit cells with values to be matched against; andselectively providing search values to said two CAM bit cells based on one or more input search values and whether the associative memory cell is in a CAM or TCAM bit mode, such that in a CAM bit mode the two CAM bit cells produce two CAM bit results, and when in a TCAM bit mode, the two CAM bit cells produce a single TCAM bit result; wherein said selectively providing search values would provide different search values to said two CAM bit cells if the associative memory was in a different CAM or TCAM bit mode for the same said one or more input search values.
26. The method of claim 25, wherein said programming of each of the two CAM bit cells includes programming a same value to each of each of the two CAM bit cells when in the TCAM bit mode and the TCAM bit represents a don't care value.
27. The method of claim 25, wherein said programming of each of the two CAM bit cells includes programming different values to said two CAM bit cells when in the TCAM bit mode and the TCAM bit represents a zero or a one.
28. An apparatus for identifying matches in an associative memory cell configured to selectively produce a matching result of two binary content-addressable memory (CAM) bits or of one ternary content-addressable memory (TCAM) bit, the associative memory cell comprising two CAM bit cells configured to identify whether or not their associated programmed value matches a respective search value; and match detection logic, coupled to said two CAM bit cells, configured to generate the matching result based on said matching of said two CAM bit cells against said search values, the apparatus comprising: means for programming each of the two CAM bit cells with values to be matched against; andmeans for selectively providing search values to said two CAM bit cells based on one or more input search values and whether the associative memory cell is in a CAM or TCAM bit mode, such that in a CAM bit mode the two CAM bit cells produce two CAM bit results, and when in a TCAM bit mode, the two CAM bit cells produce a single TCAM bit result; wherein said selectively providing search values would provide different search values to said two CAM bit cells if the associative memory was in a different CAM or TCAM bit mode for the same said one or more input search values.
29. The apparatus of claim 28, wherein said means for programming of each of the two CAM bit cells is configured to program a same value to each of each of the two CAM bit cells when in the TCAM bit mode and the TCAM bit represents a don't care value.
30. The apparatus of claim 28, wherein said means for programming of each of the two CAM bit cells is configured to program different values to said two CAM bit cells when in the TCAM bit mode and the TCAM bit represents a zero or a one.

CROSS-REFERENCE TO RELATED APPLICATIONS

This is a continuation-in-part of two applications: (a) copending nonprovisional application Ser. No. 11/252,960, now U.S. Pat. No. 7,266,004 filed Oct. 18, 2005, and of (b) Ser. No. 11/444,730, filed Jun. 1, 2006, which is a continuation-in-part of nonprovisional application Ser. No. 11/252,960, now U.S. Pat. No. 7,266,004 filed Oct. 18, 2005; with both of these applications being incorporated by reference in their entirety.

US Referenced Citations (14)

Number	Name	Date	Kind
6181698	Hariguchi	Jan 2001	B1
6307798	Ahmed et al.	Oct 2001	B1
6307855	Hariguchi	Oct 2001	B1
6362991	Barnes	Mar 2002	B1
6717946	Hariguchi et al.	Apr 2004	B1
6721202	Roge et al.	Apr 2004	B1
6731526	Inoue	May 2004	B2
6879504	Lien et al.	Apr 2005	B1
7042747	Castagnetti et al.	May 2006	B1
7266004	Krishnan et al.	Sep 2007	B2
20050010719	Slavin	Jan 2005	A1
20050289295	Shoham et al.	Dec 2005	A1
20060080498	Shoham et al.	Apr 2006	A1
20060168494	Trainin	Jul 2006	A1

Related Publications (1)

	Number	Date	Country
	20070088910 A1	Apr 2007	US

Continuation in Parts (3)

	Number	Date	Country
Parent	11444730	Jun 2006	US
Child	11498609		US
Parent	11252960	Oct 2005	US
Child	11444730		US
Parent	11252960	Oct 2005	US
Child	11252960		US

Associative memory cells configured to selectively produce binary or ternary content-addressable memory lookup results

Information

Patent Number

Date Filed

Date Issued

Inventors

Original Assignees

Examiners

Agents

CPC

US Classifications

Field of Search

US

International Classifications