CN115567590A - Data packet scheduling method, device, equipment and readable storage medium - Google Patents

Abstract

The application discloses a data packet scheduling method, a device, equipment and a readable storage medium, wherein the method comprises the following steps: acquiring an external data packet; determining a data packet to be scheduled, which accords with a preset data type, in the external data packet; matching the data packet to be scheduled with data in a preset cache table; and if the first destination address corresponding to the data packet to be scheduled in the preset cache table is obtained through matching, carrying out redirection processing on the data packet to be scheduled, and forwarding the redirected data packet to be scheduled to the first destination address. According to the method and the device, the external data packet is obtained, the data packet to be scheduled which accords with the preset data type is selected, the data packet to be scheduled is matched with the data in the preset cache table, if the first destination address corresponding to the data packet to be scheduled is obtained through matching, the data packet to be scheduled is redirected and forwarded to the first destination address, the data packet to be scheduled is prevented from being forwarded again through a user state, and therefore the scheduling performance of the data packet is improved.

Description

Data packet scheduling method, device, device and readable storage medium

technical field

The present application relates to the field of data processing, and in particular to a data packet scheduling method, device, equipment and readable storage medium.

Background technique

In the application of the network technology simulation verification platform, in order to simulate more realistic and complex scenarios, the network technology simulation verification platforms in multiple places will be linked with each other to form a combined simulation verification platform. In order to coordinate and monitor network traffic, the combined simulation verification platform often selects a main platform to coordinate and schedule other platforms. In order for the main platform to supervise and control the behavior of users and sub-platforms, all sub-platforms and user traffic need to be forwarded through the main platform.

In commonly used data scheduling solutions, scheduling strategies and algorithms are run in user mode, which leads to the fact that when the main platform forwards, all data packets not only need to pass through the complex network protocol stack in the kernel, but also need to be processed multiple times. Data copying between the user state and the kernel state allows the scheduling service in the user state to make decisions on the forwarding address of the data packet.

Due to the complexity of the network protocol stack in the kernel and the frequent data transfer between the kernel state and the user state, a lot of unnecessary machine resources are occupied, and finally the traditional data scheduling solution is used in the scenario of large data flow. Performance drops drastically.

Contents of the invention

In view of this, the present application provides a data packet scheduling method, device, device, and readable storage medium, aiming at improving the scheduling performance of data packets.

In order to achieve the above object, the present application provides a data packet scheduling method, the data packet scheduling method includes the following steps:

Get external data packets;

Determine the data packets to be scheduled that meet the preset data types in the external data packets;

Matching the data packet to be scheduled with the data in the preset cache table;

If the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is matched, then the to-be-scheduled data packet is redirected, and the redirected to-be-scheduled data packet is forwarded to the The first destination address.

Exemplarily, after matching the data packet to be scheduled with the data in the preset cache table, it includes:

If the first destination address is not matched, perform initialization processing on the to-be-scheduled data packet, and output the initialized to-be-scheduled data packet to the user state;

After formulating the scheduling policy of the initialized data packet to be scheduled in the user state, acquiring the second destination address of the scheduling policy;

updating the second destination address to the preset cache table to ensure that the initialized data packet to be scheduled matches the preset cache table to obtain the second destination address.

Exemplarily, if the first destination address is not matched, perform initialization processing on the data packet to be scheduled, including:

If the first destination address is not matched, then calculate the first hash value and the first key value group of the data packet to be scheduled;

Updating the data of the first hash value and the first key value group to the preset cache table, so as to complete the initialization process of the to-be-scheduled data packet.

Exemplarily, the updating the second destination address to the preset cache table includes:

Analyzing the scheduling policy to obtain a parsed second destination address and a second hash value;

matching a second key-value group corresponding to the second hash value in the preset cache table;

Determine the data packet to be updated based on the second key-value group;

updating the second destination address to the data packet to be updated, and updating the second destination address to the preset cache table.

Exemplarily, the updating the second destination address to the data packet to be updated, and updating the second destination address to the preset cache table include:

Based on the second hash value and the second key value group, determine whether the second destination address in the preset cache table is consistent with the data packet to be updated;

If inconsistent, update the second destination address to the data packet to be updated, and update the second destination address to the preset cache table, and mark the number of matches in the update process as one;

If they are consistent, then determine the number of matches between the second destination address and the data packet to be updated;

If the number of matches is greater than or equal to two times, then marking the data packet to be updated is valid;

If the matching times are less than two times, then add one matching times.

Exemplarily, the matching the data packet to be scheduled with the data in the preset cache table includes:

Calculate key-value data of the data packet to be scheduled: the key-value data includes protocol, source port, destination port and source address;

combining at least two data in the key-value data to obtain the retrieval key-value;

The retrieval key value is matched with the data in the preset cache table.

Exemplarily, if the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is obtained by matching, redirecting the to-be-scheduled data packet includes:

If the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is matched, then the to-be-scheduled data packet is transmitted to the constructor, so that the constructor uses the first destination address update to the data packet to be scheduled, and update the source port and source address of the data packet to be scheduled by using the current node data.

Exemplarily, in order to achieve the above purpose, the present application also provides a data packet scheduling device, the device comprising:

An acquisition module, used to acquire external data packets;

A determining module, configured to determine a data packet to be scheduled in the external data packet based on a preset data type;

A matching module, configured to match the data packet to be scheduled with the data in the preset cache table;

A judging module, configured to perform redirection processing on the data packet to be scheduled and forward the redirected data packet to be scheduled if the first destination address corresponding to the data packet to be scheduled in the preset cache table is matched. The data packet is sent to the first destination address.

Exemplarily, in order to achieve the above purpose, the present application also provides a data packet scheduling device, the device includes: a memory, a processor, and a data packet scheduler stored in the memory and operable on the processor , the data packet scheduling program is configured to implement the steps of the data packet scheduling method as described above.

Exemplarily, in order to achieve the above purpose, the present application also provides a computer-readable storage medium, on which a data packet scheduler is stored, and when the data packet scheduler is executed by a processor, the above-mentioned The steps of the data packet scheduling method described above.

Compared with the existing technology, multiple data copies are performed between the kernel state and the user state to obtain decisions such as the scheduling strategy and forwarding address made in the user state. The kernel state forwards the data packet according to the decision data, resulting in Compared with the situation where the performance of data packet scheduling is low, in this application, the external data packets are obtained, and according to the preset data type, the data packets that meet the preset data type are determined from the external data packets as the data packets to be scheduled, and the The data packet to be scheduled is matched with the data in the preset cache table, and if the first destination address corresponding to the data packet to be scheduled in the preset cache table is matched, the data packet to be scheduled is redirected, and Forward the redirected data packet to be scheduled to the first destination address, that is, directly determine the first destination address of the data packet to be scheduled in the kernel state, and redirect the data packet to be scheduled, so that the data packet to be scheduled can be directly The data packet to be scheduled is forwarded to the first destination address, thereby avoiding obtaining scheduling decision data from the user state, and also avoiding the process of multiple data copies between the user state and the kernel state. Therefore, in this application, improving The scheduling performance of data packets is improved.

Description of drawings

Fig. 1 is a schematic flow chart of the first embodiment of the data packet scheduling method of the present application;

FIG. 2 is a schematic flow diagram of a second embodiment of the packet scheduling method of the present application;

FIG. 3 is a data update flow chart of the second embodiment of the data packet scheduling method of the present application;

FIG. 4 is a schematic diagram of a third embodiment of the packet scheduling method of the present application;

FIG. 5 is a schematic structural diagram of the hardware operating environment involved in the solution of the embodiment of the present application.

The realization, functional features and advantages of the present application will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

detailed description

It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application.

The present application provides a data packet scheduling method. Referring to FIG. 1 , FIG. 1 is a schematic flowchart of a first embodiment of the data packet scheduling method of the present application.

The embodiment of the present application provides an embodiment of the data packet scheduling method. It should be noted that although the logic sequence is shown in the flow chart, in some cases, the sequence shown or described steps. For the convenience of description, the execution subject describes each step of the data packet scheduling method below, and the data packet scheduling method includes:

Step S110: Acquiring external data packets;

In the application of the network technology simulation verification platform, the main platform will forward the traffic corresponding to the sub-platform and the user mode through the main platform, so that the main platform can assist and monitor the traffic of the sub-platform, but the scheduling strategy and algorithm are in the user In the state, when the main platform is forwarded, the data packet needs to pass through the network protocol stack in the kernel state and reach the user state, and the data packet is copied multiple times between the user state and the kernel state in order to obtain the user state Data such as the scheduling policy and the destination address of the scheduling.

External data packets are data packets that need to be scheduled through the kernel mode.

The acquisition process is to acquire the external data packet from the XDP (eXpress Data Path, fast data path) hook (hook).

Step S120: Determine the data packets to be scheduled that meet the preset data type in the external data packets;

External data packets are forwarded through the kernel state, and are divided into two forwarding modes according to the actual forwarding situation. One is that the traditional data packets are copied multiple times in the kernel state and user state, and finally forwarded to the destination address by the kernel state. One is that in this embodiment, the destination address of the data packet is determined through the kernel state, and the data packet is directly forwarded in the kernel state.

Exemplarily, in the way of directly forwarding data packets in the kernel mode, it is necessary to determine the data type of the data packets first, and select the data packets that are directly forwarded in the kernel mode. All other packets use the traditional forwarding mode.

Among them, the number of external data packets is huge, and the data types of the external data packets cover various types, such as: video, audio and so on. For actual application scenarios, when forwarding different types of data, the amount of data is emphasized. For example, in the scenario of forwarding video data packets and audio data packets, the number of video data packets exceeds 70% or 80% of the total external data packets. etc. (the weight ratio determined according to the actual situation), the forwarding data of the current scene focuses on the forwarding of video data packets. Therefore, according to this situation, the external data packets can be filtered and directly forwarded to the destination address using the kernel mode. The video data packets are processed accordingly, while the audio and video are still forwarded in the way of multiple copies between the kernel state and the user state.

At the same time, all the data types included in the external data packet can also be directly forwarded in the kernel state. Similarly, the data packet to be scheduled can also be selected with a preset data type, which can be any data in the data packet. One type, or multiple types, and the preset data type depends on the actual usage scenario and purpose.

When determining the data packet to be scheduled, the rx filter can be used to analyze the data packet in the rx filter, and filter according to the analyzed quintuple to find the data packet related to the user mode. The matching rules of the rx filter can be flexibly configured according to user needs. Including but not limited to supporting filtering based on ports, protocols, etc.

Step S130: matching the data packet to be scheduled with the data in the preset cache table;

Exemplarily, cached data is stored using per-cpu (per-cpu variable) Hash (hash) in eBPF (extended Berkeley Packet Filter, extended Berkeley packet filter). In order to avoid cached data from occupying too many kernel resources, e -schedule (scheduling function) uses the LRU (Least Recently Used, page replacement algorithm) strategy to replace the cached content. The cache content of the e-schedule is stored in the preset cache table. The preset cache table includes two hash tables, namely redirect_cache_Map (mapping table of source address and destination address correspondence) in Table 1, and key_content_Map (business The mapping table of the corresponding relationship between the hash value of the data and the source address). The specific structure is shown in Table 1 and 2:

Table 1 Mapping table of source address and destination address correspondence

Table 2 Mapping table of the correspondence between the hash value of business data and the source address

Table 1 stores the corresponding relationship between the source address and the redirection address. Corresponding to the matching domain calculation function, the cache key (keyword) is composed of protocol, source, destination port and source IP (address). Because the value (value) in the cache, the key-value data composed of key-value needs to be composed of in-link and out-link data packets, so the valid field is set to indicate whether the current data is available to avoid misuse of the e-schedule. Full cache data.

Table 2 stores the corresponding relationship between the business data hash and the key (keyword, which is the source address and other data in Table 1) in Table 1.

In the cache item initialization function, this embodiment will record all data except the destination port/IP (destination address), and store it in the corresponding cache table.

Match the data packet to be scheduled with the preset cache table. The matching process is mainly carried out based on the format of the stored data in the preset cache table. According to the relevant data of the data packet, the corresponding The destination address, the destination address is the destination forwarding address of the data packet, that is, the kernel mode needs to forward the data packet to the destination address.

The matching process is to determine the whereabouts of the data packets from the preset cache table.

Exemplarily, the matching the data packet to be scheduled with the data in the preset cache table includes:

Step a: Calculate the key-value data of the data packet to be scheduled: the key-value data includes protocol, source port, destination port and source address;

Table 1 in the preset cache table is the correspondence between the source address (key value) and the redirected destination address, and Table 2 is the correspondence between the hash value of the business data and the key value in Table 1. When the data packet is matched with the preset cache table, the key-value data in the data packet needs to be calculated first.

Exemplarily, the key-value data includes protocol, source port, destination port and source address, and the key-value data comes according to the five-tuple of the data packet to be scheduled, and the five-tuple data of the data packet includes: source address, source port, A set of five quantities consisting of destination address, destination port, and protocol can be used to determine the key-value data of the data packet by analyzing the quintuple data of the data packet to be scheduled.

The above parsing process can be parsed in a filter for filtering data packets conforming to a preset data type.

Step b: combining at least two data in the key-value data to obtain the retrieval key-value;

The key-value data includes four kinds of data, and the above-mentioned four kinds of data are randomly combined, and the combination process can be arbitrarily combined according to at least two data in the key-value data, so as to obtain the retrieval key value.

Exemplarily, the data of the retrieval key can be any combination of two or any three data, for example, the combination of the source address and the source port is used as the retrieval key, and for example, the destination port and the source port are used as the retrieval key , and for example, the random combination of source address, source port and protocol as the search key.

Step c: matching the retrieval key value with the data in the preset cache table.

Use the search key value to match in the preset cache table, and match the search key value with the data in the preset cache table. If the match is successful, the destination address can be determined in the preset cache table. If the match fails, then Unable to get the destination address from the preset cache table.

Step S140: If the first destination address corresponding to the data packet to be scheduled in the preset cache table is matched, perform redirection processing on the data packet to be scheduled, and forward the redirected data packet to be scheduled to the first destination address.

The first destination address is the destination address stored in the preset cache table when the preset cache table is updated.

The traditional forwarding mode is to copy data between the kernel state and the user state multiple times, so that the kernel state obtains the forwarding strategy of the user state. In this embodiment, when the first destination address is obtained from the preset cache table, the second A destination address redirects the data packet to be scheduled, directly changes the destination address of the data packet to be scheduled in the kernel state, and directly forwards the redirected data packet to be scheduled to the first destination address.

In this redirection process, the quintuple data of the data packet to be scheduled is adjusted, and the destination address and source address are changed, so that the kernel state directly forwards the data packet to be scheduled to the destination address, thereby avoiding the data packet to be scheduled to pass through The network protocol stack and the multiple copy process between the user state and the kernel state, thereby improving the packet scheduling performance.

Exemplarily, if the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is obtained by matching, redirecting the to-be-scheduled data packet includes:

Step d: If the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is matched, then the to-be-scheduled data packet is transmitted to the constructor for the constructor to use the first destination address A destination address is updated to the data packet to be scheduled, and a source port and a source address of the data packet to be scheduled are updated using current node data.

When the first destination address is matched, the quintuple data of the data packet to be scheduled needs to be changed, so that the data packet to be scheduled can be forwarded directly through the kernel state.

In order to change the data packet to be scheduled, transmit the data packet to the constructor, update the first destination address to the data packet to be scheduled through the constructor, and update the current node data to the source port and source address of the data packet to be scheduled , so that the kernel state directly forwards the data packets to be scheduled, and avoids the need to go through the network protocol stack. By copying data between the kernel state and the user state multiple times, the data related to the scheduling policy is obtained from the user state, and the data packets to be scheduled are The forwarding process is performed, thereby shortening the forwarding process and improving the scheduling performance of data packets.

Compared with the existing technology, multiple data copies are performed between the kernel state and the user state to obtain decisions such as the scheduling strategy and forwarding address made in the user state. The kernel state forwards the data packet according to the decision data, resulting in Compared with the situation where the performance of data packet scheduling is low, in this application, the external data packets are obtained, and according to the preset data type, the data packets that meet the preset data type are determined from the external data packets as the data packets to be scheduled, and the The data packet to be scheduled is matched with the data in the preset cache table, and if the first destination address corresponding to the data packet to be scheduled in the preset cache table is matched, the data packet to be scheduled is redirected, and Forward the redirected data packet to be scheduled to the first destination address, that is, directly determine the first destination address of the data packet to be scheduled in the kernel state, and redirect the data packet to be scheduled, so that the data packet to be scheduled can be directly The data packet to be scheduled is forwarded to the first destination address, thereby avoiding obtaining scheduling decision data from the user state, and also avoiding the process of multiple data copies between the user state and the kernel state. Therefore, in this application, improving The scheduling performance of data packets is improved.

Exemplarily, referring to FIG. 2, FIG. 2 is a schematic flowchart of the second embodiment of the packet scheduling method of the present application. Based on the first embodiment of the packet scheduling method of the present application, a second embodiment is proposed, and the method further includes:

Step S210: If the first destination address is not matched, perform initialization processing on the data packet to be scheduled, and output the initialized data packet to be scheduled to the user state;

In the process of matching the data packet to be scheduled with the preset cache table, there is a situation that the data packet to be scheduled is not included in the preset cache table, therefore, in the matching process, it is impossible to determine the first A destination address, that is, the data packet to be scheduled cannot be directly redirected through the kernel mode, and the data packet cannot be directly forwarded either.

At this time, according to the traditional forwarding mode, it is necessary to perform information interaction between the kernel state and the user state, so as to obtain the scheduling decision data of the data packet to be scheduled in the user state, and forward the scheduling data packet according to the decision data. Therefore, it is necessary to transmit the data packets to be scheduled to the user state through the network protocol stack, and obtain corresponding scheduling decision data from the user state.

In order to avoid encountering such data packets that do not save the key value and destination address in the preset cache table next time, a self-learning planning mode is set up, and the scheduling decision-making mechanism of the user state is self-learned in the kernel state. Scheduling decision-making mechanism, record the corresponding scheduling record to the preset cache table, so that when the next data packet to be scheduled matches the preset cache table, the destination address can be obtained directly in the kernel state, avoiding the same type of data to be scheduled When the package is packaged, data copies between the user state and the kernel state are performed multiple times.

Before the data packet to be scheduled is output to the user state, the data packet to be scheduled is initialized. The initialization process, that is, the data packet to be scheduled is processed according to the traditional forwarding mode, and at the same time, the data packet to be scheduled is processed. Key-value data, hash values of business data and other data are updated to the preset cache table. After the scheduling decision data is obtained, the preset cache table is updated.

Exemplarily, the initialization process includes storing corresponding data in a preset cache table, and at the same time, includes performing corresponding processing on the data packets to be scheduled according to the traditional forwarding mode.

Exemplarily, if the first destination address is not matched, perform initialization processing on the data packet to be scheduled, including:

Step e: If the first destination address is not matched, then calculate the first hash value and the first key value group of the data packet to be scheduled;

When the matching fails to obtain the first destination address, it is necessary to store relevant data such as key value data and service data of the data packet to be scheduled in the preset cache table, and the storage format of the preset cache table includes the key value of the data packet, For related data such as destination address and service data, it is necessary to calculate the first hash value and the first key-value group of the data packet to be scheduled.

Among them, the first hash value is the relevant data value of the service data, and the first key-value group is the key-value data of the data packet to be scheduled, including four data items of protocol, source port, destination port and source address, and the calculated data After that, fill in the corresponding preset cache table.

Step f: updating the data of the first hash value and the first key value group to the preset cache table, so as to complete the initialization process of the data packet to be scheduled.

In the process of updating data to the preset cache table, update the first hash value to Table 2, update the first key-value group to Table 1 and Table 2, and update the remaining corresponding parameters except the destination port and destination address to the preset cache table, so as to complete the initialization processing of the data packets to be scheduled.

Step S220: After formulating the scheduling strategy for the initialized data packets to be scheduled in the user state, obtain the second destination address of the scheduling strategy;

After the data packets to be scheduled are initialized and output to the user state, the user state will formulate a corresponding scheduling policy according to the data packets to be scheduled, and include scheduling decision data such as the second destination address and related business data in the scheduling policy.

At the same time, when obtaining the scheduling policy, the corresponding data packet to be scheduled will be obtained (the process of copying the data packet multiple times in the traditional forwarding mode). When obtaining the data packet, the tx filter is used to filter the data packet. The process is the same as the above-mentioned rx filter screening process, and the filtering method is also the same, and will not be repeated here.

After formulating the scheduling strategy for the initialized data packets to be scheduled in the user state, the second destination address contained in the scheduling strategy is obtained.

Step S230: updating the second destination address to the preset cache table to ensure that the initialized data packet to be scheduled matches the preset cache table to obtain the second destination address.

When the data packet to be scheduled is initialized, the key value and hash value in the data packet are stored in the preset cache table, and in order to schedule the data packet to be scheduled, the destination address of the data packet needs to be determined, Therefore, when the second destination address of the scheduling policy is obtained from the user state, the second destination address is updated into the preset cache table, so that in the preset cache table, the key value and The data corresponding to the destination address is completed, and then when the data packets to be scheduled are subsequently scheduled, the matching of the destination addresses can be completed directly according to the preset cache table, and the redirection and forwarding of the data packets are directly performed by the kernel state.

Exemplarily, the updating the second destination address to the preset cache table includes:

Step g: Analyzing the scheduling policy to obtain the parsed second destination address and second hash value;

In updating the second destination address to the preset cache table, it is necessary to determine the update position of the second destination address, that is, to determine the correspondence between the second destination address and the data packets to be scheduled in the preset cache table, therefore, Before updating the second destination address to the preset cache table, the scheduling strategy needs to be parsed to calculate the second destination address and the second hash value of the scheduling strategy from the user state. The second hash value is also Hash value of business data.

Step h: matching the second key-value group corresponding to the second hash value in the preset cache table;

In the current preset cache linked list, the first hash value and the first key-value group data of the data packet to be scheduled are saved. When using the second hash value and the preset cache table for matching, if the result is matched, it is determined in The data packet to be scheduled corresponding to the scheduling policy in the user mode is matched in the preset cache table, and the result of the matching is the second key-value group corresponding to the second hash value in the preset cache table, according to the logic of data update However, the scheduling policy given by the user mode is the second hash value and the second destination address for the data packet to be scheduled, and the second hash value and the matched second key-value group should both be the first hash value value and the data corresponding to the first destination address, so as to complement the data correspondence relationship of the data packet to be scheduled in the preset cache table.

Step i: Determine the data package to be updated based on the second key-value group;

The first hash value and the first key value group in the current preset cache table determine the corresponding data packet to be scheduled, and after determining the corresponding relationship with the data packet to be scheduled according to the second hash value and the second key value group, The preset cache table and the data packets to be scheduled can be updated.

The to-be-updated data packet is a to-be-scheduled data packet of to-be-updated data.

Step j: updating the second destination address to the data packet to be updated, and updating the second destination address to the preset cache table.

During the update process, in addition to updating the second destination address to the data packet to be updated, at the same time, it is necessary to update the second destination address to the preset cache table.

Since the first hash value and the second hash value, the first key value group and the second key value group are all related data of the data packet to be scheduled, that is, the above hash value and key value group are in the preset cache table It represents the same type of data packets to be scheduled, so when updating the preset cache table, only the second destination address should be updated to the corresponding position.

Exemplarily, the updating the second destination address to the data packet to be updated, and updating the second destination address to the preset cache table include:

Step k: Based on the second hash value and the second key value group, determine whether the second destination address in the preset cache table is consistent with the data packet to be updated;

Before updating the data, it is necessary to determine whether the data in the scheduling policy corresponds to the data in the preset cache table one by one to ensure the consistency of the data and avoid the mismatch between data packets and the key values in the preset cache table, which may cause There is an error in kernel mode forwarding packets.

According to the second hash value and the second key value group, determine the corresponding position of the data of the first hash value and the second key value group in the preset cache table, and after the data of both parties are successfully matched, the second purpose can be determined The address is consistent with the packet to be updated.

Exemplarily, the current process of determining whether the data is consistent is the process of matching the data in the scheduling policy with the data in the preset cache table and the data in the data packet to be updated. If the data is directly determined to be consistent, then it is determined that the data to be updated If there is already a scheduling strategy formulated by the user state in the data packet and the preset cache table, but they are inconsistent, it is determined that there is no scheduling strategy in the data packet to be updated and the preset cache table.

When the matching determination data is consistent, it is determined that the current scheduling strategy has been updated to the preset cache table, and it is necessary to further determine the validity of the relevant data of the current scheduling strategy.

When the matching determines that the data is inconsistent, it is determined that there is no relevant data of the scheduling strategy in the current preset cache table, and then the relevant data in the scheduling strategy needs to be updated to the preset cache table.

Wherein, the relevant data of the scheduling policy includes data such as hash value, key-value group, and destination address.

Step 1: If inconsistent, update the second destination address to the data packet to be updated, and update the second destination address to the preset cache table, and mark the number of matches in the update process as one;

When the matching results are inconsistent, the second destination address is updated to the data packet to be updated and the preset cache table. At the same time, for subsequent judgment of the validity of the data, after the update is completed, the number of matches marked in the update process is once.

When the matching results are consistent, it proves that the relevant data of the current scheduling strategy has been stored in the preset cache table and the data packet to be updated, but the number of matches of the data in the preset cache table is one, and the data is invalid data. It cannot be called directly, therefore, it will need to transfer the scheduled data packets to the user state.

Therefore, when the matching results are consistent, it is necessary to further judge the matching times, so as to determine the validity of the relevant data.

Step m: If they are consistent, determine the number of matches between the second destination address and the data packet to be updated;

Step n: if the number of matches is greater than or equal to two times, mark the data packet to be updated as valid;

Step o: If the number of matching times is less than two times, increase the number of matching times by one time.

If the second destination address is consistent with the data packet to be updated, it proves that the second destination address can be applied to the forwarding of the current data packet to be scheduled, but in order to ensure the accuracy of the data, when the matching is successful, the number of matching times needs to be determined.

Exemplarily, the user mode will not modify the forwarded service data. Therefore, for the problem that the five-tuple of the forwarded data packet (the data packet to be scheduled processed by the kernel mode) and the original data packet (as the data packet stored in the preset cache table) are different, the hash value of the business data is used to find the forwarding data packet. The packet matches the original packet.

There may be a hash value conflict problem in this mechanism. Therefore, in the case of possible conflicts, in this embodiment, the data packets of the matching flow are required to be successfully matched twice in a row (Match_count>=2, the number of matches is greater than or equal to 2 ) to make this cache entry valid.

However, when the number of matching times is less than two times, the validity of the data cannot be ensured. Therefore, when the number of matching times is less than two times, the number of matching times is increased once.

Referring to FIG. 3, FIG. 3 is a flow chart of data updating in this embodiment.

In Figure 3, after obtaining the relevant data of the scheduling policy, it is filtered through the tx filter. The filtering process is the same as the content described above, and will not be repeated here. The process includes parsing the scheduling policy and the data packets transmitted from the user mode, calculating the second hash value of the scheduling policy, and determining the second key-value group from the preset cache table according to the second hash value, and according to the second The key-value group determines the data packet to be updated from the preset cache table, and matches the relevant data of the scheduling policy with the data in the preset cache table and the data of the data packet to be updated to determine whether they are consistent. When the data is inconsistent, Update the preset cache table and the data package to be updated, and mark the number of matches in the update process as one, and when the data is consistent, further judge whether the number of matches is greater than or equal to two, and if the number of matches is greater than or equal to two , that is, the updated data is valid, and when the number of matches is less than twice, add one time to the current number of matches, so that when the number of matches is subsequently judged, the subsequent number of matches is guaranteed to meet the validity judgment and valid data It can be used in the preset cache table, and the data whose matching times do not exceed two times is invalid data. When the data packet to be scheduled is matched in the kernel state, this type of data existing in the preset cache table is not used.

In this embodiment, when the matching is to obtain the first destination address, the data packet to be scheduled is initialized, the relevant data of the data packet to be scheduled is stored in the preset cache table, and the data packet to be scheduled is initialized The packet is output to the user mode, and after the user mode formulates a scheduling policy according to the data packet to be scheduled, the second destination address in the scheduling policy is obtained, and the second destination address is stored in the preset cache table, so as to ensure that When the subsequent kernel state processes this type of data packet to be scheduled, it matches the second destination address in the preset cache table, and then achieves the process of self-learning in the kernel state for the scheduling policy of the user state, avoiding subsequent forwarding of this type of data packet to be scheduled The process of multiple data copies between the kernel state and the user state is required. Therefore, the scheduling strategy of the user state is recorded in a self-learning manner to improve the performance of subsequent data packet scheduling.

Exemplarily, referring to FIG. 4, FIG. 4 is a schematic flow chart of the third embodiment of the data packet scheduling method of the present application. Based on the above-mentioned first embodiment and the second embodiment of the data packet scheduling method of the present application, a third embodiment is proposed:

Figure 4 includes the data processing process and data interaction process between the user state and the kernel state.

The user state contains relevant data for formulating scheduling policies. Therefore, there is a data scheduling system in the user state, and the scheduling policies obtained by the kernel state from the user state are all formulated by the data scheduling system.

There is a network protocol stack between the user state and the kernel state as an information exchange guarantee, and the kernel state is directly connected to the network device.

There are two data processing modules in the kernel state, namely the in-chain processing module and the out-chain processing module, and different data processing operations are performed in the two modules.

The data processing process of the in-chain module in the kernel mode is as follows:

After the kernel state obtains the external data packet through the XDP hook, the external data packet passes through the rx filter to filter out some data types used for direct forwarding in the kernel state, and performs matching domain calculation on the filtered data packet, that is, the matching data packet and The data in the preset cache table, after the destination address is matched, the data packet is transmitted to the constructor, and the data packet is constructed and redirected. When the matching calculation fails to match the relevant destination address, it will MISS is reported, that is, the matching is not successful (when the matching is successful, HIT will be reported), that is, it proves that the relevant data of the data packet does not exist in the preset cache table. At this time, the data packet is initialized and passed through the network protocol stack. , output the data packet to the data scheduling system in user mode.

The data processing process of the out-of-chain module in the kernel mode is as follows:

After the data scheduling system in the user mode formulates the scheduling policy of the data packet, the scheduling policy is obtained through the tc hook. The obtaining process also includes obtaining relevant data packets and other data. At this time, the data is filtered through the tx filter. The filtering process is the same as that of the rx filter, which filters out specific types of data packets. After filtering, the scheduling policy is analyzed to determine the destination address, hash value and other data, and according to the data and the preset cache The corresponding data item is updated according to the matching result of the table, and the data update includes data such as a preset cache table and data packets to be scheduled.

In this embodiment, through the in-chain processing module, data packets are directly forwarded in the kernel state, and through the out-of-chain processing module, the kernel state performs self-learning of the scheduling strategy of the user state, so that the kernel state saves more scheduling strategies, To deal with different data packets, avoid multiple data copies between user mode and kernel mode.

In addition, the present application also provides a data packet scheduling device, and the data packet scheduling device includes:

An acquisition module, used to acquire external data packets;

A determining module, configured to determine a data packet to be scheduled in the external data packet based on a preset data type;

A matching module, configured to match the data packet to be scheduled with the data in the preset cache table;

A judging module, configured to perform redirection processing on the data packet to be scheduled and forward the redirected data packet to be scheduled if the first destination address corresponding to the data packet to be scheduled in the preset cache table is matched. The data packet is sent to the first destination address.

Exemplarily, the matching module includes:

The first judging submodule is used to initialize the data packet to be scheduled if the first destination address is not matched, and output the initialized data packet to be scheduled to the user state;

The obtaining submodule is used to obtain the second destination address of the scheduling policy after the scheduling policy of the initialized data packet to be scheduled is formulated in the user state;

The update submodule is configured to update the second destination address to the preset cache table, so as to ensure that the initialized data packet to be scheduled matches the preset cache table to obtain the second destination address.

Exemplarily, the judging submodule includes:

A judging unit, configured to calculate the first hash value and the first key value group of the data packet to be scheduled if the first destination address is not matched;

A first update unit, configured to update the first hash value and the data of the first key value group to the preset cache table, so as to complete the initialization process of the data packet to be scheduled.

Exemplarily, the update submodule includes:

A parsing unit, configured to parse the scheduling policy to obtain a parsed second destination address and a second hash value;

a matching unit, configured to match a second key-value group corresponding to the second hash value in the preset cache table;

a determining unit, configured to determine the data package to be updated based on the second key-value group;

A second updating unit, configured to update the second destination address to the data packet to be updated, and update the second destination address to the preset cache table.

Exemplarily, the second updating unit includes:

A determination subunit, configured to determine whether the second destination address in the preset cache table is consistent with the data packet to be updated based on the second hash value and the second key value group;

The first judging subunit is configured to update the second destination address to the data packet to be updated if inconsistent, update the second destination address to the preset cache table, and mark the number of matches in the update process for one time;

The second judging subunit is configured to determine the number of matches between the second destination address and the data packet to be updated if they are consistent;

A third judging subunit, configured to mark the data packet to be updated as valid if the number of matches is greater than or equal to two;

The fourth judging subunit is configured to increase the matching times by one if the matching times are less than two times.

Exemplarily, the matching module includes:

A calculation submodule, configured to calculate key-value data of the data packet to be scheduled: the key-value data includes protocol, source port, destination port and source address;

A combination submodule, configured to combine at least two data in the key-value data to obtain a retrieval key value;

The matching submodule is used to match the retrieval key value with the data in the preset cache table.

Exemplarily, the judgment module includes:

The second judging submodule is configured to transmit the to-be-scheduled data packet to the constructor if the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is matched, for the construction The function uses the first destination address to update the to-be-scheduled data packet, and uses current node data to update the source port and source address of the to-be-scheduled data packet.

The specific implementation manners of the data packet scheduling device of the present application are basically the same as the embodiments of the above data packet scheduling method, and will not be repeated here.

In addition, the present application also provides a data packet scheduling device. As shown in FIG. 5 , FIG. 5 is a schematic structural diagram of a hardware operating environment involved in the solution of the embodiment of the present application.

Exemplarily, FIG. 5 is a schematic structural diagram of a hardware operating environment of a data packet scheduling device.

As shown in FIG. 5, the data packet scheduling device may include a processor 501, a communication interface 502, a memory 503, and a communication bus 504, wherein the processor 501, the communication interface 502, and the memory 503 complete mutual communication through the communication bus 504, The memory 503 is used to store computer programs; the processor 501 is used to implement the steps of the data packet scheduling method when executing the programs stored in the memory 503 .

The communication bus 504 mentioned in the above data packet scheduling device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (Extended Industry Standard Architecture, EISA) bus or the like. The communication bus 504 can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

The communication interface 502 is used for communication between the data packet scheduling device and other devices.

The memory 503 may include a random access memory (Random Access Memory, RMD), and may also include a non-volatile memory (Non-Volatile Memory, NM), such as at least one disk memory. Optionally, the memory 503 may also be at least one storage device located away from the aforementioned processor 501 .

The above-mentioned processor 501 may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; it may also be a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

The specific implementation manner of the data packet scheduling device of the present application is basically the same as the embodiments of the above data packet scheduling method, and will not be repeated here.

In addition, the embodiment of the present application also proposes a computer-readable storage medium, where a data packet scheduler is stored on the computer-readable storage medium, and when the data packet scheduler is executed by a processor, the above-mentioned data packet schedule is implemented. method steps.

The specific implementation manners of the computer-readable storage medium of the present application are basically the same as the embodiments of the data packet scheduling method described above, and will not be repeated here.

It should be noted that, as used herein, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or system comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, method, article, or system. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article or system comprising that element.

The serial numbers of the above embodiments of the present application are for description only, and do not represent the advantages and disadvantages of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware, but in many cases the former is better implementation. Based on such an understanding, the technical solution of the present application can be embodied in the form of a software product in essence or the part that contributes to the prior art, and the computer software product is stored in a storage medium as described above (such as ROM/RAM , magnetic disk, optical disk), including several instructions to enable a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in various embodiments of the present application.

The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structures or equivalent process transformations made by using the description of the application and the accompanying drawings are directly or indirectly used in other related technical fields. , are all included in the patent protection scope of the present application in the same way.

Claims (10)

1. A packet scheduling method, characterized in that, the packet scheduling method comprises the following steps: Get external data packets; Determine the data packets to be scheduled that meet the preset data types in the external data packets; Matching the data packet to be scheduled with the data in the preset cache table; If the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is matched, then the to-be-scheduled data packet is redirected, and the redirected to-be-scheduled data packet is forwarded to the The first destination address. 2. The data packet scheduling method according to claim 1, wherein, after matching the data packet to be scheduled with the data in the preset cache table, the method includes: If the first destination address is not matched, perform initialization processing on the to-be-scheduled data packet, and output the initialized to-be-scheduled data packet to the user state; After formulating the scheduling policy of the initialized data packet to be scheduled in the user state, acquiring the second destination address of the scheduling policy; updating the second destination address to the preset cache table to ensure that the initialized data packet to be scheduled matches the preset cache table to obtain the second destination address. 3. The data packet scheduling method according to claim 2, wherein if the first destination address is obtained without matching, then the data packet to be scheduled is initialized, including: If the first destination address is not matched, then calculate the first hash value and the first key value group of the data packet to be scheduled; Updating the data of the first hash value and the first key value group to the preset cache table, so as to complete the initialization process of the to-be-scheduled data packet. 4. The data packet scheduling method according to claim 3, wherein said updating said second destination address to said preset cache table comprises: Analyzing the scheduling policy to obtain a parsed second destination address and a second hash value; matching a second key-value group corresponding to the second hash value in the preset cache table; Determine the data packet to be updated based on the second key-value group; updating the second destination address to the data packet to be updated, and updating the second destination address to the preset cache table. 5. The data packet scheduling method according to claim 4, wherein the updating the second destination address to the data packet to be updated, and updating the second destination address to the preset cache table ,include: Based on the second hash value and the second key value group, determine whether the second destination address in the preset cache table is consistent with the data packet to be updated; If inconsistent, update the second destination address to the data packet to be updated, and update the second destination address to the preset cache table, and mark the number of matches in the update process as one; If they are consistent, then determine the number of matches between the second destination address and the data packet to be updated; If the number of matches is greater than or equal to two times, then marking the data packet to be updated is valid; If the matching times are less than two times, then add one matching times. 6. The packet scheduling method according to claim 1, wherein said matching said packet to be scheduled with data in a preset cache table comprises: Calculate key-value data of the data packet to be scheduled: the key-value data includes protocol, source port, destination port and source address; combining at least two data in the key-value data to obtain the retrieval key-value; The retrieval key value is matched with the data in the preset cache table. 7. The data packet scheduling method according to claim 6, wherein if the first destination address corresponding to the data packet to be scheduled in the preset cache table is obtained by matching, then the data packet to be scheduled is Packets are redirected, including: If the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is matched, then the to-be-scheduled data packet is transmitted to the constructor, so that the constructor uses the first destination address update to the data packet to be scheduled, and update the source port and source address of the data packet to be scheduled by using the current node data. 8. A data packet scheduling device, characterized in that, the data packet scheduling device comprises: An acquisition module, used to acquire external data packets; A determining module, configured to determine a data packet to be scheduled in the external data packet based on a preset data type; A matching module, configured to match the data packet to be scheduled with the data in the preset cache table; A judging module, configured to perform redirection processing on the to-be-scheduled data packet and forward the redirected to-be-scheduled data packet if the first destination address corresponding to the to-be-scheduled data packet in the preset cache table is matched. The data packet is sent to the first destination address. 9. A data packet scheduling device, characterized in that the device includes: a memory, a processor, and a data packet scheduler stored on the memory and operable on the processor, the data packet scheduler It is configured to realize the steps of the data packet scheduling method according to any one of claims 1 to 7. 10. A computer-readable storage medium, characterized in that, a data packet scheduler is stored on the computer-readable storage medium, and when the data packet scheduler is executed by a processor, any one of claims 1 to 7 is implemented. The steps of the packet scheduling method described in item.

Images