WO2016090848A1

WO2016090848A1 - Method and device for managing table space of ternary content addressable memory

Info

Publication number: WO2016090848A1
Application number: PCT/CN2015/079107
Authority: WO
Inventors: 姜海明
Original assignee: 中兴通讯股份有限公司
Priority date: 2014-12-11
Filing date: 2015-05-15
Publication date: 2016-06-16
Also published as: CN105743678A

Abstract

Provided are a method and device for managing the table space of a ternary content addressable memory (TCAM), the method comprising: configuring a first parameter of a first register of the TCAM, the first parameter comprising a lookup-table space scope, the lookup-table space scope corresponding to a plurality of blocks; configuring a second parameter of a second register of the TCAM according to the lookup-table space scope, the second parameter comprising the space size of each of the blocks within the lookup-table space scope and the width of the lookup-table; writing the table in the adjacent first block and second block, and the table writing direction of the first block is opposite to that of the second block; determining if the first block and the second block are fully written; if the first block is fully written but the second block is not, adjusting the first parameter and the second parameter, and writing the table according to the adjusted first parameter and second parameter. The method of the present invention enables a TCAM to realize hardware block capacity sharing.

Description

Method and device for managing tri-state content addressing memory TCAM table space

Technical field

The present invention relates to the field of communications technologies, and in particular, to a method and apparatus for managing a TCAM table space of a tri-state content addressed memory.

Background technique

With the development of the network, the growth of network traffic and the emergence of new services, network equipment needs to have wire speed and flexible processing capabilities. Currently, the network chip includes two types: Application Specific Integrated Circuits (ASICs) and Network Processors (NPs). With its high-speed processing and flexible programmability, network processors have become an effective solution for data processing in today's networks.

There are many different memories in the network processor, such as Static Random Access Memory (SRAM), Ternary Content Addressable Memory (TCAM), Synchronous Dynamic Random Memory (Synchronous Dynamic). Random Access Memory (SDRAM), etc., these memories store various service items, such as port table, hardware address (Media Access Control, MAC) table, routing table, and so on.

The TCAM is mainly used to quickly search for an Access Control List (ACL) and routing entries. The main feature of TCAM is fast parallel search. The application on the network processor is usually that the microcode composes a key value from the message extraction information, and sends a search command to the TCAM through the NP TCAM controller, and processes the message according to the result.

TCAM is characterized by high cost, high power consumption and limited capacity. The advantages are fast search, fixed delay, longest prefix matching and multi-region matching. In order to make full use of the valuable TCAM space, TCAM generally adopts multi-table distinguishing technology on hardware. A TCAM is divided into blocks of several sizes, and the key value sent by the microcode is searched in parallel in multiple blocks at the same time. Send key values at a time, returning multiple lookup results.

In order to support different widths of the search, each block must be configured with a lookup table width W. For example, the NL11K series of NETLOGIC has a typical width of 80, 160, 320, 640; IDT's TCAM has a typical width of 72, 144, 288, and 576.

For example, an Internet Protocol version 4 (IPv4)/Internet Protocol version 6 (IPv6) ACL, the IPv4 quintuple width is 107 bytes (bit); the IPv6 ACL matches the IPv6 quintuple width to 299 bits. With the NL11K series, the corresponding database of IPv4 should be configured according to the 160-bit width, and the database type should be classified and matched; the corresponding block of IPv6 ACL should be configured according to the 320-bit width. Due to the different widths, these two service entries must be placed in different hardware blocks.

At present, the common practice is that the width of the lookup table of each block is determined to be dead, and the write table in each block is sequentially written from the address low to the high address. For a typical application, IPv4 ACL and IPv6 ACL are allocated in two adjacent hardware blocks. The allocated space is 64K*80bit, and the table width is 160bit and 320bit respectively. The IPv4 ACL has a capacity of 32K and the IPv6 ACL has a capacity of 16K. The downside of this approach is that capacity sharing is not possible.

Summary of the invention

It is an object of embodiments of the present invention to provide a method and apparatus for managing a TCAM table space of a tri-state content addressed memory, which enables TCAM to implement hardware block capacity sharing.

In order to achieve the above object, an embodiment of the present invention provides a method for managing a TCAM table space of a tri-state content addressed memory, the method comprising:

Configuring a first parameter of the first register of the TCAM, the first parameter includes a table space range, and the table space range corresponds to a plurality of block blocks;

Configuring a second parameter of the second register of the TCAM according to the table space range, where the second parameter includes a space size and a table width of each block in the table space;

Writing a table in the adjacent first block and the second block, and the writing direction of the first block is opposite to the writing direction of the second block;

Determining whether the first block and the second block are full, if the first block is full and the second block is not full, adjusting the first parameter and the second parameter to select the portion of the second block adjacent to the first block The space is assigned to the first block, and the table is written according to the adjusted first parameter and the second parameter.

Among them, the method also includes:

If the first block and the second block are not full, continue to write the table;

If both the first block and the second block are full, the information that failed to write the table is returned.

The step of writing a table in the adjacent first block and the second block, and the writing direction of the first block is opposite to the writing direction of the second block includes:

The table is written in the first block and the second block in the direction from the address low to the address high and the address high to the address low, respectively.

Wherein, determining whether the first block and the second block are full, if the first block is full and the second block is not full, adjusting the first parameter and the second parameter, and following the adjusted first parameter and the second parameter Before the steps of writing the table, the method further includes:

Detecting the presence of fragments in the second block;

If there is fragmentation, the fragment is compressed.

The IPv4 access control list ACL is written in the first block, and the lookup table width of the first block is 160 bits.

The second block is written with an IPv6 ACL, and the second block has a table width of 320 bits.

Embodiments of the present invention also provide an apparatus for managing a TSTAT table space of a tri-state content addressed memory, the apparatus comprising:

a first configuration module, configured to configure a first parameter of the first register of the TCAM, the first parameter includes a table space range, and the table space range corresponds to a plurality of block blocks;

a second configuration module, configured to configure a second parameter of the second register of the TCAM according to the table space range, where the second parameter includes a space size and a table width of each block in the table space;

Writing a table module, configured to write a table in the adjacent first block and the second block, and the writing direction of the first block is opposite to the writing direction of the second block;

The determining module is configured to determine whether the first block and the second block are full, and trigger the adjustment module when the first block is full and the second block is not full.

The adjusting module is configured to adjust the first parameter and the second parameter according to the triggering of the determining module, to allocate the partial space adjacent to the first block in the second block to the first block, and trigger the writing table module to be adjusted according to the adjustment The first parameter and the second parameter are written in the table.

Wherein, the device further comprises:

The determining module triggers the write table module to continue to write the table when the first block and the second block are not full, and triggers the return module when both the first block and the second block are full;

Return to the module, set to return the information that failed to write the table according to the trigger of the judgment module.

Among them, the write table module includes:

The first unit is configured to write the table in the first block and the second block in a direction from the address low to the address high and the address high to the address low, respectively.

Wherein, the device further comprises:

a detecting module, configured to detect whether there is a fragment in the second block, and trigger the compression module when there is fragmentation;

The compression module is set to compress the fragments.

The above aspects of the present invention include at least the following beneficial effects:

In the method for managing the TCAM table space of the tri-state content addressing memory in the embodiment of the present invention, two services of different table look widths are respectively placed in the adjacent first block and the second block, and the first block and the first block are respectively The write direction of the second block is reversed, so that when the table in the first block is full and the table in the second block is full, the parameters of the first register and the second register can be configured to The adjacent space of a block is allocated to the first block, so that the TCAM realizes hardware block capacity sharing.

DRAWINGS

1 is a flow chart showing the steps of a method for managing a TCAM table space of a tri-state content addressable memory according to an embodiment of the present invention;

2 is a schematic diagram showing the space of the first block and the second block before dynamically adjusting the spatial size of the first block and the second block in the embodiment of the present invention;

3 is a schematic diagram showing the space of the first block and the second block after dynamically adjusting the spatial size of the first block and the second block in the embodiment of the present invention;

FIG. 4 is a schematic structural diagram of an apparatus for managing a TCAM table space of a tri-state content addressable memory according to an embodiment of the present invention.

detailed description

In order to make the technical problems, technical solutions, and advantages of the present invention more apparent, the following detailed description will be made in conjunction with the accompanying drawings and preferred embodiments.

The embodiment of the present invention provides a method and a device for managing a TCAM table space of a tri-state content-addressable memory in the prior art that the TCAM hardware block cannot share capacity, and enables the TCAM to implement hardware block capacity sharing.

As shown in FIG. 1, an embodiment of the present invention provides a method for managing a TCAM table space of a tri-state content addressed memory, the method comprising:

Step 11: Configure a first parameter of the first register of the TCAM, where the first parameter includes a table space range, and the table space range corresponds to multiple block blocks.

In a preferred embodiment of the present invention, the first parameter of the first register is used to indicate a table lookup space range when the table is looked up by the TCAM, and the lookup table range corresponds to a plurality of hardware blocks.

Step 12: Configure a second parameter of the second register of the TCAM according to the table space range, where the second parameter includes a space size and a table width of each block in the table space.

In a preferred embodiment of the present invention, the lookup table width and space size of the tables placed in the respective hardware blocks of the TCAM may be different. For example, in the current network, the IPv6 ACL is not configured much, and the IPv4 ACL is configured more. Therefore, the space for the hardware block for the IPv6 ACL is smaller than the space for the hardware block for the IPv4 ACL. At the same time, since the width of the IPv4 quintuple is 107 bits and the width of the IPv6 quintuple is 299 bits, the lookup width of the hardware block used for placing the IPv4 ACL is different from the width of the lookup table of the hardware block for placing the IPv6 ACL. The preferred look-up table width will be configured according to the model of the TCAM. For example, the NL11K series, the corresponding database of IPv4 should be configured according to the 160-bit width, and the corresponding block of the IPv6 ACL should be configured according to the 320-bit width.

Step 13, writing a table in the adjacent first block and the second block, and the write direction of the first block is opposite to the write direction of the second block, preferably in a direction from a low address to an address high, respectively. The address is written in the first block and the second block in the direction from the address high to the low address.

In a preferred embodiment of the invention, the first block entry is written from address low to address high, and the second block entry is written from address high to address low, such that when the first block is full and the second block is not full The space size of the first block and the second block can be re-allocated by adjusting the first parameter and the second parameter. In addition, in the embodiment of the present invention, the two adjacent blocks can be configured to write the table according to the above-mentioned write table method to implement capacity sharing, and the other blocks in the TCAM can still be written in the traditional write table manner. .

Step 14: Determine whether the first block and the second block are full. If the first block is full and the second block is not full, adjust the first parameter and the second parameter to compare the first block with the first block. The partial space of the neighbor is assigned to the first block, and the table is written according to the adjusted first parameter and the second parameter.

In the above embodiment of the present invention, the method further includes: determining whether the first block and the second block are full, and if the first block and the second block are not full, continuing to write the table; if the first block and the first block When both blocks are full, the information that failed to write the table is returned.

In the above embodiment of the present invention, it is determined in step 14 whether the first block and the second block are full, and if the first block is full and the second block is not full, the first parameter and the second parameter are adjusted. And before the step of writing the table according to the adjusted first parameter and the second parameter, the method further includes: detecting whether there is fragment in the second block; if there is fragmentation, compressing the fragment, so that the software management is simple, not A large number of entries are involved, and the forwarding surface is not affected.

In a preferred embodiment of the present invention, the reserved capacity of the first block and the second block may be configured, for example, one third of the first block and the second block are reserved, and the reserved space is reserved for itself. enjoy. In this way, the service in the first block does not occupy two block spaces, and the user later configures the service of the second block without any space available.

In the preferred embodiment of the present invention, the ACL table is taken as an example. Since the user occasionally performs an ACL deletion operation, there may be free space in the TCAM hardware block. Of course, it is possible that each deletion operation is accompanied by a compression operation, and It may not be compressed and wait until the appropriate time for compression, depending on the TCAM table management algorithm. The function of the compression is to concentrate the scattered fragments in the second block into a continuous space, and move the concentrated continuous space to a position adjacent to the first block, so that the first block occupies the fragment space.

Optionally, in the foregoing embodiment of the present invention, the IPv4 access control list ACL is written in the first block, and the lookup table width of the first block is 160 bits; the second block is written with an IPv6 ACL, and the second block is The lookup table width is 320bit.

In a preferred embodiment of the present invention, a preferred example is taken as an example. After configuring the first parameter of the first register and the second parameter of the second register, writing in the first block in the direction from the address low to the address high IPv4 ACL, in the second block, writes the IPv6 ACL in the direction from the address high to the low address, as shown in Figure 2. Before dynamically adjusting the space size of the first block and the second block, the space of the first block is M, and the second The space of the block is N. When it is detected that the first block is full and the second block is not full, the space size of the first block and the second block can be dynamically adjusted. That is, the first parameter and the second parameter are adjusted, and the space P adjacent to the first block in the second block is assigned to the first block, and the lookup width of the space P is also changed to be the same as the lookup width of the space M. Thus, as shown in FIG. 3, the space of the first block is expanded to M+N, and the space of the second block is reduced to NP. The first register and the second register need to be determined according to the specific model of the TCAM. For example, if the TCAM of the NETLOGIC NL series is used in this example, the second register includes a Block Select Register (BSR) in the Logical Table Register (LTR) group, and a block. Configuration Register (Block Configuration Register, BCR for short).

In order to achieve the above objective, as shown in FIG. 4, an embodiment of the present invention further provides an apparatus for managing a TCAM table space of a tri-state content addressed memory, the apparatus comprising:

The first configuration module 41 is configured to configure a first parameter of the first register of the TCAM, where the first parameter includes a table space range, and the table space range corresponds to multiple block blocks;

The second configuration module 42 is configured to configure a second parameter of the second register of the TCAM according to the table space range, where the second parameter includes a space size and a table width of each block in the table space;

The write table module 43 is configured to write a table in the adjacent first block and the second block, and the write table direction of the first block is opposite to the write table direction of the second block;

The determining module 44 is configured to determine whether the first block and the second block are full.

And when the first block is full and the second block is not full, the adjustment module 45 is triggered,

The adjusting module 45 is configured to adjust the first parameter and the second parameter according to the triggering of the determining module 44, to allocate the partial space adjacent to the first block in the second block to the first block, and trigger the writing table module 43 Write the table according to the adjusted first parameter and second parameter.

Wherein, the device further comprises:

The determining module 44 triggers the write table module 43 to continue writing the table when both the first block and the second block are not full, and triggers the return module when both the first block and the second block are full;

The write table module 43 includes:

Wherein, the device further comprises:

The detecting module is configured to detect whether there is a fragment in the second block, and when there is fragmentation, trigger the compression module,

The compression module is set to compress the fragments.

Industrial Applicability: As can be seen from the above description, the present invention puts two different look-up-width services into adjacent first blocks and second blocks, and the writing directions of the first block and the second block are opposite, such that When the table in the first block is full and the table in the second block is full, the part of the second block adjacent to the first block can be classified by configuring the parameters of the first register and the second register. Give the first block, so that the TCAM implements hardware block capacity sharing.

The above is a preferred embodiment of the present invention, and it should be noted that those skilled in the art can also make several improvements and retouchings without departing from the principles of the present invention. It should be considered as the scope of protection of the present invention.

Claims

A method of managing a TSTAT table space of a tri-state content addressed memory, comprising:

Configuring a first parameter of the first register of the TCAM, the first parameter includes a table space range, and the table space range corresponds to a plurality of block blocks;

And configuring, according to the table space range, a second parameter of the second register of the TCAM, where the second parameter includes a space size and a table look width of each block in the table lookup space;

Writing a table in the adjacent first block and the second block, and the write table direction of the first block is opposite to the write table direction of the second block;

Determining whether the first block and the second block are full, if the first block is full and the second block is not full, adjusting the first parameter and the second parameter to A partial space adjacent to the first block in the second block is allocated to the first block, and the table is written according to the adjusted first parameter and the second parameter.
The method of claim 1 wherein the method further comprises:

If the first block and the second block are not full, continue to write the table;

If the first block and the second block are both full, the information that the write table failed is returned.
The method of claim 1, wherein said writing a table in adjacent first block and second block, and said write direction of said first block is opposite to a write direction of said second block The steps include:

The table is written in the first block and the second block in a direction from an address low to an address high and an address high to an address low, respectively.
The method of claim 1, wherein the determining whether the first block and the second block are full, if the first block is full and the second block is not full, then adjusting the method Before the step of writing the first parameter and the second parameter according to the adjusted first parameter and the second parameter, the method further includes:

Detecting whether there is fragment in the second block;

If there is fragmentation, the fragment is compressed.
The method of claim 1, wherein the first block is written with an Internet Protocol version 4 IPv4 access control list ACL, and the lookup table width of the first block is 160 byte bits.
The method of claim 1, wherein the second block is written with an Internet Protocol version 6 IPv6 access control list ACL, and the lookup width of the second block is 320 bits.
An apparatus for managing a TSTAT table space of a tri-state content addressed memory, comprising:

a first configuration module, configured to configure a first parameter of the first register of the TCAM, the first parameter includes a lookup table space range, and the lookup table space range corresponds to a plurality of block blocks;

a second configuration module, configured to configure a second parameter of the second register of the TCAM according to the table space range, where the second parameter includes a space size and a check of each block in the table lookup space Table width

Writing a table module, configured to write a table in the adjacent first block and the second block, and the write table direction of the first block is opposite to the write table direction of the second block;

a determining module, configured to determine whether the first block and the second block are full, and triggering an adjustment module when the first block is full and the second block is not full

The adjusting module is configured to adjust the first parameter and the second parameter according to the triggering of the determining module, to assign a part of the second block adjacent to the first block to the first Block, and triggering the write table module to write the table according to the adjusted first parameter and the second parameter.
The device of claim 7 wherein said device further comprises:

The determining module triggers the write table module to continue writing the table when the first block and the second block are not full, and when the first block and the second block are full, Trigger return module;

The return module is configured to return information that the write table failed according to the trigger of the determining module.
The apparatus of claim 7 wherein said write table module comprises:

The first unit is configured to write a table in the first block and the second block in a direction from an address low to an address high and an address high to an address low, respectively.
The device of claim 7 wherein said device further comprises:

a detecting module, configured to detect whether there is a fragment in the second block, and trigger a compression module when there is fragmentation;

A compression module configured to compress the fragments.
The apparatus of claim 7, wherein the first block is written with an IPv4 access control list ACL, and the lookup width of the first block is 160 bits.
The apparatus of claim 7, wherein the second block is written with an IPv6 ACL, and the second block has a look-up table width of 320 bits.