KR102172056B1 - Control method, apparatus and program of token-based caching system including icn router and content provider terminal - Google Patents

Abstract

Provided is a token-based caching system. The method of controlling a token-based caching system including an ICN router and a content provider terminal includes the steps of: determining, by the ICN router, a token generation rate based on a cache hit rate; determining, by the ICN router, the total amount of tokens to be allocated to the content provider terminal based on the determined token generation rate and the amount of unmarked content; determining, by the content provider, a marking speed for controlling the number of contents to be marked among the plurality of contents; determining, by the content provider terminal, first content marked for caching in the ICN router and second content unmarked not for caching in the ICN router, based on the marking speed and content popularity; transmitting, by the content provider terminal, the first content and the second content to the ICN router; and subtracting, by the ICN router, the amount of tokens corresponding to the received first content from the total amount of tokens of the content provider terminal, and caching the first content in the ICN router.

Description

Control method, device and program of token-based caching system including ICN router and content provider terminal {CONTROL METHOD, APPARATUS AND PROGRAM OF TOKEN-BASED CACHING SYSTEM INCLUDING ICN ROUTER AND CONTENT PROVIDER TERMINAL}

The present invention relates to a method, apparatus, and program for controlling a token-based caching system including an ICN router and a content provider terminal in an information-centric network (ICN).

With the development of Internet technology, information-oriented network (Information Centric Networking) that transmits data through name information of content, beyond TCP/IP technology that transmits data between two nodes based on location information such as IP address. Research on ICN) technology is ongoing.

Specifically, information-oriented network (ICN) technology can transmit data through name information of content, not through location information of a terminal. In addition, data copies are stored in nodes on the network during the process of transmitting data (in-network caching), so users can receive services by using copies of contents located closer to each other. There is an effect to experience the response time.

In ICN caching technology, performance can be determined by caching policies such as whether to store data in a location and whether to replace data, and furthermore, if the integrity of the cached data is not guaranteed, malicious users can There is also a security issue in which content contaminated by can be served. Therefore, the ICN router must perform additional functions in relation to the optimal caching policy and security issues.

Routers in the existing TCP/IP-based network perform routing with IP addresses of a fixed size, while routers in ICN perform routing functions under a much larger namespace (content names of varying lengths) and various caching related Since functions are performed, overhead is likely to occur, and as a result, limited performance may be degraded due to a bottleneck of the router.

Conventional technologies for solving the above problems include Hashing and Bloom filters, and Cisco is developing an ICN prototype router capable of routing at a speed of 20 Gbps or higher. Optimization techniques are being disclosed. However, since the above-described various methods aim to optimize gains due to caching, it is a reality that they do not solve the problem of overhead added to the router.

Therefore, there is a need for a method capable of maximizing caching efficiency while reducing overhead due to caching.

The problem to be solved by the present invention is to provide a method, apparatus, and program for controlling a token-based caching system including an ICN router and a content provider terminal.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems that are not mentioned will be clearly understood by those skilled in the art from the following description.

A method for controlling a token-based caching system including an ICN router and a content provider terminal according to an aspect of the present invention for solving the above-described problem includes the steps of: determining, by the ICN router, a token generation rate based on a cache hit ratio; Determining, by the ICN router, a total amount of tokens to be allocated to the content provider terminal based on the determined token generation rate and the amount of unmarked content; Determining, by the content provider, a marking speed for controlling the number of contents to be marked among a plurality of contents; Determining, by the content provider terminal, a first content marked for caching in the ICN router and a second content not marked for caching in the ICN router, based on the marking speed and content popularity; Transmitting, by the content provider terminal, the first content and the second content to the ICN router; And subtracting, by the ICN router, the amount of tokens corresponding to the received first content from the total amount of tokens of the content provider terminal, and caching the first content to the ICN router.

Other specific details of the present invention are included in the detailed description and drawings.

According to various embodiments of the present invention described above, an ICN router is an effective method for maximizing a cache hit ratio.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a system diagram illustrating a token-based caching system according to an embodiment of the present invention.
2 is a flowchart illustrating the operation of a token-based caching system according to an embodiment of the present invention.
3 is a graph for explaining a normalized frequency of access to content provided by different content provider terminals according to an embodiment of the present invention.
4 to 22 are exemplary diagrams showing experimental data for evaluating the performance of a token-based caching system according to an embodiment of the present invention.
23 is a block diagram of an apparatus according to an embodiment of the present invention.

Advantages and features of the present invention, and a method of achieving them will become apparent with reference to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in a variety of different forms, only the present embodiments are intended to complete the disclosure of the present invention, It is provided to fully inform the technician of the scope of the present invention, and the present invention is only defined by the scope of the claims.

The terms used in the present specification are for describing exemplary embodiments and are not intended to limit the present invention. In this specification, the singular form also includes the plural form unless specifically stated in the phrase. As used in the specification, “comprises” and/or “comprising” do not exclude the presence or addition of one or more other elements other than the mentioned elements. Throughout the specification, the same reference numerals refer to the same elements, and “and/or” includes each and all combinations of one or more of the mentioned elements. Although "first", "second", and the like are used to describe various elements, it goes without saying that these elements are not limited by these terms. These terms are only used to distinguish one component from another component. Therefore, it goes without saying that the first component mentioned below may be the second component within the technical idea of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used in the present specification may be used as meanings that can be commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not interpreted ideally or excessively unless explicitly defined specifically.

The terms "unit" or "module" used in the specification means software, hardware components such as FPGA or ASIC, and "unit" or "module" performs certain roles. However, "unit" or "module" is not meant to be limited to software or hardware. The "unit" or "module" may be configured to be in an addressable storage medium or may be configured to reproduce one or more processors. Thus, as an example, "sub" or "module" refers to components such as software components, object-oriented software components, class components and task components, processes, functions, properties, It includes procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays and variables. Components and functions provided within "sub" or "module" may be combined into a smaller number of components and "sub" or "module" or into additional components and "sub" or "module" Can be further separated.

Spatially relative terms "below", "beneath", "lower", "above", "upper", etc., as shown in the figure It can be used to easily describe the correlation between a component and other components. Spatially relative terms should be understood as terms including different directions of components during use or operation in addition to the directions shown in the drawings. For example, if a component shown in a drawing is turned over, a component described as "below" or "beneath" of another component will be placed "above" the other component. I can. Accordingly, the exemplary term “below” may include both directions below and above. Components may be oriented in other directions, and thus spatially relative terms may be interpreted according to orientation.

In the present specification, a computer refers to all kinds of hardware devices including at least one processor, and may be understood as encompassing a software configuration operating in a corresponding hardware device according to embodiments. For example, the computer may be understood as including all of a smartphone, a tablet PC, a desktop, a laptop, and a user client and an application running on each device, but is not limited thereto.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Each of the steps described herein is described as being performed by a computer, but the subject of each step is not limited thereto, and at least some of the steps may be performed by different devices according to embodiments.

1 is a system diagram illustrating a token-based caching system according to an embodiment of the present invention.

As shown in FIG. 1, the token-based caching system may include a content provider terminal 10, an ICN router 20, and a user terminal 30.

The content provider terminal 10 may determine content to be cached or marked on the ICN router 20 according to a preset criterion. In one embodiment, the content provider terminal 10 may determine the content to be cached or marked based on the popularity of the content.

In the present specification, marking is to add a specific index to content or designate a value of a specific data set. The content provider terminal 10 may mark the content to be cached, and the ICN router may cache the marked content.

In an embodiment, the marking may be set by designating a value in a specific field (eg, a ch field) of a data packet transmitted by the content provider terminal 10. For example, when the value of the ch field is 1, the ICN router 20 may determine that the corresponding content is marked, and when the value of the ch field is 0, the ICN router 20 may determine that the corresponding content is not marked.

The ICN router 20 may perform tasks related to routing and caching. The caching operation performed by the ICN router 20 may include searching for cached contents and checking/writing contents for the cache.

In addition, the ICN router 20 may determine the amount of tokens to be allocated to the content provider terminal 20. The amount of tokens allocated to the content provider may be obtained by Equation 1 below.

[Equation 1]

In this case, T may be an amount of tokens allocated to the content provider terminal 10, γ may be a token generation rate, and B may be unmarked content (ie, content in which a value of the ch field is set to 0).

At this time, the token generation rate can be obtained based on a cache hit rate. That is, the ICN router 20 can control the amount of content actually inserted into the cache by using the token generation rate. That is, if the ICN router 20 generates too small a token, a situation in which really popular content cannot be put into the cache may occur, resulting in a decrease in caching efficiency. Conversely, if the ICN router 20 generates too many tokens, caching a lot of content unnecessarily increases caching overhead, and at the same time, caching gains may decrease as unpopular content is stored in the cache. have. Accordingly, the ICN router 20 can control the total amount of tokens to be provided to the content provider terminal 10 by adjusting the token generation rate.

Meanwhile, according to an embodiment, when content is inserted into a cache, it may be managed by a cache replacement policy such as LRU, LFU, and MFU. Content that is popular enough to remain in the cache according to their respective policies may be considered delegated to the router by the content provider and may not be related to token usage until removed from the cache. In other words, the token is deducted from the total token when the content is first cached, and until it is removed from the cache, the cached content does not affect the amount of tokens.

That is, the ICN server 20 allocates a token to be provided to the content provider terminal 10 based on the ratio of the marked content and the unmarked content among the content received from the content provider terminal 10, It is possible to prevent (10) from indiscriminately making a content caching request, thereby preventing caching overhead of the router.

In one embodiment, it may be assumed that the first content provider transmits 11 contents to the ICN router 20 and the second content provider provides 110 contents to the ICN router 20. When the token generation rate is 0.1, the first content provider may mark up to 1 content and cache it in the ICN router 20, and the second content provider may mark up to 10 content and cache it in the ICN router 20. can do.

The user terminal 30 is a terminal for requesting content from the content provider terminal 10. When the content requested by the user terminal 30 is cached in the ICN router, the cache hit ratio may increase, and vice versa, the cache hit ratio may decrease.

2 is a flowchart illustrating the operation of a token-based caching system according to an embodiment of the present invention.

)를 결정할 수 있다.In step S110, the ICN router 20, based on the cache hit rate, the token generation rate (

) Can be determined.

In an embodiment, when the cache hit rate is less than or equal to a preset value, the ICN router 20 may decrease the value of the token generation rate, and when the cache hit rate exceeds a preset value, the ICN router 20 may increase the value of the token generation rate. However, the present invention is not limited thereto, and the ICN router 20 increases the value of the token generation rate when the cache hit rate is less than or equal to the preset value, and decreases the value of the token generation rate when the cache hit rate exceeds the preset value. Of course you can.

In one embodiment, the token generation rate may be determined by Algorithm 1 below.

Specifically, as shown in FIG. 3, the graph of FIG. 3A and the graph of FIG. 3B are normalized access frequencies for the content provided by the first content provider (hereinafter, CP1) and the second content provider (hereinafter, CP2), respectively. Represents. At this time, the rank of content on the x-axis may mean a ranking for the popularity of each content. In other words, the more popular content, the lower the ranking.

Specifically, FIG. 3A shows a result of a situation in which the same amount of user requests are generated for contents of CP1 and CP2. Although the same amount of user requests are generated for the contents of CP1 and CP2, it can be seen that the distribution of popularity of CP1 and CP2 is different. In this case, the content provided by CP1 and CP2 may follow a Zipf distribution function having parameters of 0.7 and 1.3, respectively.

3B is a graph showing a result of a case where the contents of CP1 and CP2 display the same popularity, but the user request for the CP1 content is 5 times larger than the user request for the CP2 content. Both results indicate that content with a lower ranking of CP1 (A1) may be requested more often than content with a higher ranking of CP2 (A2). In the case of FIG. 3B, if a value of a token generation rate that is too small is used, more requested content (A1) cannot be inserted into the cache, so the cache hit rate may drop. Furthermore, there are cases in which it is difficult to quickly update the cache as necessary because sufficient contents are not inserted into the cache when a value of the token generation rate that is too small is used.

Accordingly, the ICN router 20 may dynamically control the value of the token generation rate in an AIAD (Additive Increase and Additive Decrease) method. In general, the ICN router 20 may gradually decrease the value of the token generation rate in order to filter out less popular content. A smaller amount of tokens may be generated due to a decrease in the value of the token generation rate and the actual content insertion rate in the ICN router 20 may decrease. When the value of the token generation rate is less than the optimal value, the popular content (which should be cached of course) cannot be inserted into the cache, and thus the cache hit rate of the ICN router 20 may decrease. In this case, the ICN router 20 may increase the value of the token generation rate until the cache hit rate increases again. That is, the ICN router 20 may determine the value of the token generation rate according to the frequency of occurrence of a cache hit rate and a cache miss rate, as shown in Algorithm 1.

In step S120, the ICN router 20 may determine the total amount of tokens to be allocated to the content provider terminal 10 based on the determined token generation rate and the amount of unmarked content.

As described above, the ICN router 20 may determine the total amount of tokens to be provided to the content provider terminal 10 based on Equation 1.

In step S130, the content provider terminal 10 may determine a marking speed δ for adjusting the number of contents to be marked among a plurality of contents.

In one embodiment, the content provider terminal 10 aims to cache as much content as possible in the ICN router 20. However, in the token-based caching system according to the present invention, since content caching occurs by marking, the content provider terminal 10 has a marking rate determined by the token generation rate.

Specifically, when the marking rate is less than the token generation rate, the content provider terminal 10 has a token capable of caching more content in the ICN router 20, but cannot fully use the token. Due to this lack, all marked content is not cached in the ICN router 20. Furthermore, since the content marked by Equation 1 does not contribute to token generation, the marking speed and the token generation speed must be equivalent in order to maximize caching efficiency. Accordingly, there is a need for the content provider terminal 10 to acquire the token generation rate and adjust the marking rate to be equivalent to the token generation rate.

The content provider terminal 10 may use at least one of a notification-based approach and an estimation-based approach as a method for determining the marking speed. A detailed description of the notification-based approach and the estimation-based approach will be described later.

In step S140, the content provider terminal 10, based on the marking speed and content popularity, the first content marked for caching in the ICN router, and the second content not marked in order not to be cached in the ICN router 20. Can be determined.

In step S150, the content provider terminal 10 may transmit the first content and the second content to the ICN router 20.

In step S160, the ICN router 20 may subtract the amount of tokens corresponding to the received first content from the total amount of tokens of the content provider terminal, and cache the first content in the ICN router 20.

Meanwhile, as described above, the content provider terminal 10 acquires or estimates the token generation rate generated by the ICN router 20 using at least one of a notification-based approach and an estimation-based approach, and obtains or estimates The marking rate can be determined based on the generated token generation rate.

In one embodiment, the content provider terminal 10 may obtain a token generation rate generated by the ICN router 20 based on a notification-based approach. That is, the content provider terminal 10 may determine the marking rate by receiving data on the token generation rate from the ICN router.

)를 계산하거나 2) 패킷에 기록 된 유효 토큰 생성 속도 값을 자신의 것을 반영하도록 업데이트 하여 토큰 생성 속도를 획득할 수 있다.However, in a general ICN environment, a plurality of routers exist between the content provider terminal 10 and the user terminal 30. Therefore, all ICN routers on the path 1) include their own token generation rate value in the interest packet, so that the content provider terminal 10 is at the effective token generation rate (

) Or 2) update the effective token generation rate value recorded in the packet to reflect your own to obtain the token generation rate.

That is, the content provider terminal 10 receives data on the token generation rate from all ICN routers on the path between the content provider terminal 10 and the user terminal 30, and based on the received data, the effective token generation rate Can be obtained. That is, in the present specification, the effective token generation rate is actually a plurality of ICN routers existing between the content provider terminal 10 and the user terminal 30, but the content provider terminal 10 and the user terminal 30 It may mean the rate of token generation when it is assumed that there is one ICN router.

In this case, the effective token generation rate may be determined based on Equation 2 below.

[Equation 2]

In this case, β may be an arbitrary constant of 0 or more and 1 or less.

However, in this case, since each of the plurality of ICN routers must obtain an effective token generation rate based on their token generation rate and transmit it to the next ICN router on the path, the size of the packet transmitted between routers increases as the number of routers on the path increases. Is increased, and each router recalculates its own effective token generation rate based on the effective token generation rate of each router received from the router on the path, thereby increasing the amount of calculation in the ICN router. However, since the effective token generation rate is calculated by the ICN router 20 and transmitted to the content provider terminal 10, the content provider terminal 10 has the advantage of determining the marking rate by acquiring the effective token generation rate without wasting any other resources. There is this.

In another embodiment, the content provider terminal 10 may obtain a token generation rate generated by the ICN router 20 based on an estimation-based approach. Specifically, the content provider terminal 10 may estimate the effective token generation rate based on the request receiving rate rather than receiving the effective token generation rate from the ICN router 20.

Specifically, the content provider terminal 10 may increase the marking speed from the initial marking speed to cache more contents in the ICN router 20. As the marking speed increases, the cached content increases, so that the request reception ratio for each content may decrease. That is, more user requests are provided by the cache of the ICN router 20, and the ratio of the request reception speed of the content provider terminal 10 may decrease.

As the marking speed increases, if the marking speed is greater than the effective token generation speed, even if the content provider terminal 10 marks the content, the allocated token is insufficient, so that the caching speed of the content may decrease. In this case, the content provider terminal 10 receives more requests even if the marking speed increases.

When a request for more than a preset number of times is received while the marking speed increases, the content provider terminal 10 determines that the marking speed is greater than the effective token generation speed, and may decrease the marking speed.

When the marking rate becomes smaller than the effective token generation rate again, the amount of cached content is limited by the marking rate, and the request reception rate increases in proportion to the decrease in the marking rate.

That is, when the request reception speed is greater than or equal to a preset speed while the marking speed decreases, the content provider terminal 10 may increase the marking speed.

Specifically, the content provider terminal 10 may determine the marking speed based on Algorithm 2 below.

On the other hand, according to various embodiments of the present invention, if the content provider terminal 10 satisfies a preset condition, the marking speed may be increased regardless of the effective token generation speed. In one embodiment, the content provider terminal 10 may increase the marking speed when the popularity of the content increases by more than a preset range within a preset time.

In other words, there may be a case in which the popularity of the content changes and the user request for the new popular content instantaneously increases rapidly. In this case, the request reception rate at the content provider terminal 10 increases regardless of the effective token generation rate. Therefore, in order to avoid such exceptional cases when estimating the effective token generation rate, the content provider terminal 10 determines that the popularity of the content has changed when the increase in the request reception rate is more than a preset value, and You can increase the marking speed to update the cache.

Hereinafter, the present invention is experimentally implemented based on FIGS. 4 to 22, and experimental data for various effects of the present invention will be described.

The performance evaluation of the token-based caching system according to the present invention is verified through simulation and experimental studies using synthetic data and real traces. As an embodiment, the present specification may perform a simulation study using ndnSIM in a dumbbell topology, DFN and Rocketfuel, and an experimental study in a test bed network of a tree structure topology composed of nodes in which NFD is installed. In the simulation topology, the bandwidth of the core and edge links is set to 100 Mbps and 10 Mbps, respectively, and all links in the test bed network can be set to have a bandwidth of 1 Gbps.

In another embodiment, the token-based caching system is evaluated with various cache replacement policies including LRU, LFU, and FIFO, and can be compared with a caching scheme sampled at random with 1% and 10% probability. In a simulation using synthetic data, 10 content provider terminals (hereinafter, CP) provide 105 unique contents, and 10 user terminals 30 can request at a rate of 50 per second. The delay between requests in each client follows an exponential distribution of λ = 0.025, and the popularity of the content provided by each content provider terminal 20 may be determined by a Zipf-Mandelbrot distribution having a parameter value α. At this time, the value of α may be between 0.7 and 1.3. In addition, all simulations using synthetic data run for 5000 seconds and results from 1000 to 4000 seconds can be displayed. The value of β used in Equation 2 may be set to 0.3.

However, it is not limited to the above-described experimental environment, and of course, results according to various experimental environments may be derived as needed.

4 to 22 are exemplary diagrams showing experimental data for evaluating the performance of a token-based caching system according to an embodiment of the present invention.

8 is an exemplary diagram showing a dumbbell topology for static content according to an embodiment of the present invention. As shown in FIG. 8, the dumbbell topology may be implemented with 4 content provider terminals, 3 ICN routers, and 4 user terminals.

Specifically, the token-based caching method operates on the LRU cache installed in R1, and the cache size of the entire network can vary from 0.01% to 1% of the total content, and various patterns of user requests can be used.

In particular, 1) the same amount of user requests is generated for each content provider terminal, and different values of α are used in the Zipf-Mandelbrot distribution (2, 3, and 1.3, 1.0 and 0.7 for the remaining 5 content provider terminals). Types of popularity distribution can be applied, and 2) If the value of α is commonly set to 1.0, the user terminal requests content provided by 5, 3 and 2 content provider terminals at a rate of 20, 50 and 100 pieces per second. I can.

4 to 12 are diagrams for explaining the performance of static content.

4 and 5 are diagrams showing results of a cache hit ratio in a cache, a cache hit ratio, and a cache hit content ratio.

As shown in FIGS. 4B and 5B, it can be seen that all token-based caching systems according to the present invention substantially reduce caching speed to 0.05% compared to the basic LRU. Nevertheless, token-based caching systems can handle user requests before more than 99% of the content stored in the cache is removed from the cache and the cache hit rate improves. The sampling technique with a content insertion rate of 1% is similar, but it can be seen that it produces a slightly smaller cache hit rate.

6 and 7 show configurations of caches according to values of α and request rate, respectively, in an implementation of two token-based caching systems.

The configuration of the cache appears in a similar form in the two implementations, and it can be seen that the inference-based approach can operate in almost the same way as the notification-based approach. As shown in FIG. 6, when the cache size is small, a large portion of the cache is occupied by the most popular content for the content provider terminal, and in this case, α may have a value of 1.3. As the cache size increases, a content provider terminal having an α value of 1.3 may have low content sharing in the cache. This may be because the content of the content provider terminal having a low value has a high access frequency when comparing the content with the same ranking low.

Meanwhile, as illustrated in FIG. 7, it can be seen that the sharing of the cache by the content provider terminal having 100 reqs / sec decreases in proportion to the cache size. This may be the result of popular content from content provider terminals where 20 and 50 reqs/sec per second remain in the cache for a longer period of time as the cache size increases.

9 and 10 are diagrams showing the performance of a token-based caching system and a token-based caching system when combined with an LFU and FIFO.

In this case, the cache size may be set to 0.5% of the total content. In general, LFU achieves a much larger cache hit ratio than LRU, but it is limitedly used because the overhead by management is much larger.

Since the basic LFU shows quite excellent performance in terms of the cache hit ratio and the cache hit ratio ratio, when the LFU is combined with the token-based caching system according to the present invention, the cache insertion overhead can be greatly reduced to 0.05%, and caching performance can also be improved. .

In another embodiment, when the token-based caching system according to the present invention is combined with FIFO, it can be seen that more excellent performance can be achieved.

A typical FIFO can extract even the most popular content from the cache, so even if more cache is inserted, it has a lower cache hit ratio than LFU. However, when implemented together with the token-based caching system and FIFO according to the present invention, the amount of content inserted into the cache is greatly reduced, so that the inserted popular content can stay in the cache longer. That is, when the token-based caching system according to the present invention and the FIFO are implemented in combination, a high cache hit rate can be achieved with little content insertion. However, the sampling technique can reduce the caching speed to 1%, but does not improve the cache hit ratio of LFU and FIFO. This may mean that the effect of random sampling may be limited to LRU.

On the other hand, as shown in Figures 11 and 12, the two implementations of the token-based caching system according to the present invention may result in a similar cache configuration in the LFU and FIFO. This could mean that the estimate-based approach can work in a similar way to the notification-based approach.

13 and 14 are diagrams for explaining the performance of dynamic content.

As described above, FIGS. 4 to 12 are related to performance evaluation for static content that does not change in popularity. However, since there are many environments where the popularity of content changes over time, there is a need for performance evaluation for dynamic content. In this case, it is assumed that the cache size is set to 0.5% of the total content, and the content popularity ranking is changed by 2450 by adding a random number θ. That is, the i-th priority content may be i+θ ranking content. When i+θ is greater than the total number of contents (N), the ranking may be i+θ-N by modular arithmetic.

In most comparison methods, it can be seen that the cache hit ratio changes as the popularity of the content changes. In the case of an LRU (1%) sampled to show similar performance to a token-based caching system in a static content environment, it can be seen that the cache hit rate is restored after a long time (Ts). It can be determined that this is because the cache is hardly updated due to a fixed small insertion probability and old contents are filled in the cache. Conversely, the token-based caching system quickly recovers the cache hit rate by using the relationship between the token generation rate and the marking rate, so the value of the token generation rate in the ICN router increases by the above-described Algorithm 1 and Algorithm 2, and the cache is the interval of Tp. It can be refreshed every time. That is, as shown in FIG. 13B, the content provider terminal can insert more content into the cache by increasing the value of the marking speed.

Specifically, as described above, in the notification-based approach, the marking rate is set by the effective token generation rate delivered from the network, and in the estimation-based approach, the content provider terminal observes a rapid increase in the number of received requests, thereby increasing the popularity of the content. The change of can be detected.

In this experimental environment, the ICN router may stop increasing the value of the token generation rate at about 2650th, which may reduce the marking rate of the content provider terminal as shown in FIG. 13C.

Meanwhile, FIG. 14 shows the result of applying various α sets (ie, various popularity gaps), and referring to FIG. 14, it can be seen that the object of the present invention can be achieved regardless of a change in α set.

Meanwhile, according to various embodiments of the present invention, a token-based caching system may perform efficient content transmission based on a distance between a content provider terminal and a user terminal.

For efficient link usage, content that has traversed longer distances (ie, traversed more routers) may have a higher priority in the cache. The token-based caching system may provide different token generation rates for each content provider terminal.

15 is a diagram for explaining the effect of providing different token generation rates to two content provider terminals.

In this case, the ratio of the weights allocated to the two content provider terminals may vary from 1 to 7.

Specifically, FIG. 15 illustrates a result of a cache hit ratio and a content ratio of a cache in two content provider terminals having different weights. As shown in Fig. 15A, as the ratio of the weight increases, more contents from the content provider terminal having a larger weight are stored in the cache, and as a result, the ratio of the total cache hit ratio is reduced as shown in Fig. 15B. Decreases. This is because the popular content of the content provider terminal having a higher weight can be more easily inserted into the cache than the popular content of the content provider terminal having a lower weight.

16 is a diagram showing a token-based caching system implemented in an actual test bed network.

As shown in FIG. 16, the test bed network is composed of 7 router nodes and 5 end hosts, and 10 consumer or producer applications can be executed in a virtual machine of each end host. NFD is installed on all nodes and can only be used by caching router nodes in the network. The network-wide cache size is set at 0.5% of the total content, and user requests can be generated based on actual tracking on the CDN of the top 10 anonymous US websites. Specifically, of the 500 million requests recorded on the San Francisco CDN server, the first 4 million requests can be used for simulation. The temporal locality of the request can be measured using the stack distance shown in FIG. 17. The resulting graph shows a very similar shape to the empirical stack distance obtained from other traces. Requests are allocated to all consumers in a round robin fashion, each consumer generating user requests at a rate of 10 pieces per second, and it is assumed that all links in the test bed have a bandwidth of 1 Gbps.

Referring to FIG. 18, since the LRU is more likely to store more popular content in a cache closer to the consumer, the largest cache hit ratio and most of the cache hit content for all methods are at edge routers (R4 to R7 in FIG. 16). Can be observed (FIGS. 18A and 18C ). Similar to the previous simulation results, the token-based caching system reduces the cache insertion at all routers to 0.06 to 0.07% (Fig.18b), but it can be seen that it achieves a cache hit ratio that is 2 to 3 times larger than the default caching method.

At this time, when the router is located at the upper level of the topology (for example, 0.065% in R4, 0.074% in R1), it can be seen that the ratio of the number of cache insertions in the token-based caching system of the present invention to the basic LRU increases. have. That is, since the most popular content is cached at the edge, the popularity of the content reaching R1 can follow the Zipf-Mandelbrot distribution function with a relatively small alpha parameter (eg 0.7). Therefore, while R1 maximizes the cache hit ratio, more content comes in and out of the cache, and it can be confirmed that the operation of the token-based caching system according to the present invention is normally performed.

19 is a diagram for confirming the effect of β in a notification-based method by repeating an experiment using various β values.

19A to 19C show performance results after the cache converges to a normal state. Overall, it can be seen that similar results are displayed regardless of the β value. However, if the β value is too large or too small (0.1 or 0.9), it can be seen that the notification-based method indicates a longer convergence time. It can be understood that this is because the marking speed is mainly controlled by the cache of the router closest to the consumer terminal or the content provider terminal, not the entire storage space of the network. Here, the convergence time is defined as the time until 95% of the caching space is filled with cache hit content, and Fig. 19D shows the normalized convergence time using the convergence time of the estimation-based approach.

Considering the experimental results, when β has a value of 0.3 or more and 0.5 or less, it can be confirmed that the determination of the marking speed by the notification-based method is most efficient. At this time, in LRU, the closer the cache is to the consumer, the more popular content is stored, and as β decreases, the value of the token generation rate of the cache closer to the consumer terminal has a greater effect on determining the effective token generation rate. I can.

20 to 22 are diagrams for explaining performance evaluation of a token-based caching system using DFN and Australian Rocketfuel ISP topologies.

Specifically, FIG. 20A shows a DFN topology, and FIG. 20B shows an Australian Rocketfuel ISP topology. Specifically, in both topologies, 5 nodes are connected to the edge routers of all consumer terminals and generate user requests at a rate of 10 per second. A single content provider terminal is connected to a producer side edge router. The cache is installed on all routers, and the size of the network-wide cache can vary from 0.01% to 1% of the total content.

21 and 22 are diagrams for explaining a cache hit ratio for all cache sizes of DFN and Australian Rocketfuel ISP and the number of insertions in the entire network, respectively.

21 and 22, the token-based caching system according to the present invention shows the largest cache hit ratio even in a large-scale topology, but it can be seen that less than 1% of content is inserted into the cache compared to the basic LRU of all cache sizes. have. In addition, the smaller the cache size, the greater the cache hit rate, and when the cache of the network is set to store 1% of all contents as compared to the basic and 1% random sampling LRU, the token-based caching system according to the present invention is used in the DFN topology. It can be seen that the cache hit ratios of 1.76 Х and 1.07 Х are generated, respectively, and the cache hit ratios of 2.19x and 1.09x are generated in the Australian Rocketfuel ISP topology, respectively.

On the other hand, when the overall cache size of the network is reduced to store 0.01% of all content, the performance increase of the cache hit ratio in the DFN topology is 8.4x and 2.35x, respectively, and the performance increase of the cache hit ratio in the topology of the Australian Rocketfuel ISP is 9.26x and respectively. You can see that it is 2.70x.

In conclusion, the token-based caching system according to the present invention can effectively reduce the caching overhead of the router without sacrificing the caching advantage even in a real environment where the topology is complex and the router has a cache of a very small size compared to the total amount of content. can confirm.

23 is a block diagram of an apparatus according to an embodiment of the present invention.

The processor 102 may include one or more cores (not shown) and a graphic processing unit (not shown) and/or a connection path (eg, a bus) for transmitting and receiving signals with other components. .

The processor 102 according to an embodiment performs the method described with respect to FIG. 2 by executing one or more instructions stored in the memory 104.

Meanwhile, the processor 102 temporarily and/or permanently stores a signal (or data) processed inside the processor 102, a RAM (Random Access Memory, not shown) and a ROM (Read-Only Memory). , Not shown) may further include. In addition, the processor 102 may be implemented in the form of a system on chip (SoC) including at least one of a graphic processing unit, RAM, and ROM.

The memory 104 may store programs (one or more instructions) for processing and controlling the processor 102. Programs stored in the memory 104 may be divided into a plurality of modules according to functions.

The steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or a combination thereof. Software modules include Random Access Memory (RAM), Read Only Memory (ROM), Erasable Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), Flash Memory, hard disk, removable disk, CD-ROM, or It may reside on any type of computer-readable recording medium well known in the art to which the present invention pertains.

Meanwhile, the device 100 shown in FIG. 23 may be implemented with at least one of a content provider terminal 10, an ICN router 20, and a user terminal 30, and the device 100 includes a processor 102 and a memory ( 104), it goes without saying that various configurations may be included.

In one embodiment, the device 100 may further include a communication unit. The communication unit may include at least one of a WiFi chip, a Bluetooth chip, a wireless communication chip, and an NFC chip. In particular, each of the WiFi chip and the Bluetooth chip can perform communication using a WiFi method and a Bluetooth method. When a WiFi chip or a Bluetooth chip is used, various types of connection information such as an SSID and a session key are first transmitted and received, and various types of information can be transmitted and received after communication is connected using the same. The wireless communication chip refers to a chip that performs communication according to various communication standards such as IEEE, Zigbee, 3rd Generation (3G), 3rd Generation Partnership Project (3GPP), and Long Term Evolution (LTE). The NFC chip refers to a chip that operates in a Near Field Communication (NFC) method using a 13.56 MHz band among various RF-ID frequency bands such as 135 kHz, 13.56 MHz, 433 MHz, 860 to 960 MHz, and 2.45 GHz.

In another embodiment, when the device 100 is implemented as the ICN router 20, the memory 104 may store a marking inspection module, a token management module, and a first managing module. The marking inspection module may be a module for determining whether the first content received from the content provider terminal is marked. The token management module is a module capable of subtracting the amount of tokens corresponding to the first content from the total amount of tokens of the content provider terminal. When the amount of the deducted token is 0 or more, the processor 102 may cache the first content in the ICN router, and when the amount of the deducted token is negative, the first content may not be cached in the ICN router. The first managing module may determine the token generation rate through the various methods described above.

In another embodiment, when the device 100 is implemented as the content provider terminal 10, the memory 104 may include a marking module and a second management module. The marking module may determine and mark a first content for caching in an ICN router among a plurality of contents stored in the memory.

The second management module may estimate the token generation rate of the ICN router based on the various methods described above.

Components of the present invention may be implemented as a program (or application) and stored in a medium in order to be executed in combination with a computer that is hardware. Components of the present invention may be implemented as software programming or software elements, and similarly, embodiments include various algorithms implemented with a combination of data structures, processes, routines or other programming elements, including C, C++ , Java, assembler, etc. may be implemented in a programming or scripting language. Functional aspects can be implemented with an algorithm running on one or more processors.

In the above, embodiments of the present invention have been described with reference to the accompanying drawings, but those of ordinary skill in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features. You can understand. Therefore, the embodiments described above are illustrative in all respects, and should be understood as non-limiting.

10: content provider terminal
20: ICN router
30: user terminal

Claims (10)

In the control method of a token-based caching system including an ICN router and a content provider terminal,
Determining, by the ICN router, a token generation rate based on a cache hit ratio;
Determining, by the ICN router, a total amount of tokens to be allocated to the content provider terminal based on the determined token generation rate and the amount of unmarked content;
Determining, by the content provider terminal, a marking speed for controlling the number of contents to be marked among a plurality of contents;
Determining, by the content provider terminal, a first content marked for caching in the ICN router and a second content not marked for caching in the ICN router, based on the marking speed and content popularity;
Transmitting, by the content provider terminal, the first content and the second content to the ICN router; And
Subtracting, by the ICN router, the amount of tokens corresponding to the received first content from the total amount of tokens of the content provider terminal, and caching the first content to the ICN router; Control method comprising a. The method of claim 1,
The step of determining the token generation rate,
Based on the following Algorithm 1, the control method, characterized in that determining the token generation rate.

The method of claim 1,
The step of determining the marking speed,
Obtaining, by the ICN router, a valid token generation rate based on the token generation rate;
Transmitting, by the ICN router, the effective token generation rate to the content provider terminal; And
Including, by the content provider terminal, determining the marking speed so that the marking speed is equivalent to the effective token generation speed
The control method, characterized in that the effective token generation rate is determined based on the following equation.
<Equation>

At this time,

Is the effective token generation rate, β is an arbitrary constant between 0 and 1 and less. The method of claim 1,
The step of determining the marking speed,
Obtaining, by the content provider terminal, an effective token generation rate based on Algorithm 2 below; And
And determining, by the content provider terminal, the marking speed so that the marking speed is equivalent to the effective token generation speed.

Memory;
Communication department; And
In the ICN router comprising a; processor,
The memory,
Marking inspection module;
Token management module and
A first managing module; Including,
The processor,
Determine whether the first content received from the content provider terminal through the marking inspection module is marked,
Through the token management module, the amount of tokens corresponding to the first content is deducted from the total amount of tokens of the content provider terminal,
When the amount of the subtracted token is 0 or more, the first content is cached in the ICN router, and when the amount of the subtracted token is negative, the first content is not cached in the ICN router,
ICN router, characterized in that determining a token generation rate through the first management module. The method of claim 5,
The processor,
ICN router, characterized in that to determine the token generation rate based on Algorithm 1 below.

Memory;
Communication department; And
In the content provider terminal comprising a; processor,
The memory,
Marking module; And
Includes; a second management module,
The processor,
Determine and mark a first content for caching in an ICN router among a plurality of contents stored in the memory through the marking module,
Through the second management module, estimating the token generation rate of the ICN router,
A content provider terminal, characterized in that the marking speed is determined based on the estimated token generation speed. The method of claim 7,
The processor,
Receiving the effective token generation rate obtained based on the token generation rate from the ICN router through the communication unit,
Determining the marking speed so that the marking speed is equivalent to the effective token generation speed,
The content provider terminal, characterized in that the effective token generation rate is determined based on the following equation.
<Equation>

At this time,

Is the effective token generation rate, β is an arbitrary constant between 0 and 1 and less. The method of claim 7,
The processor,
Based on Algorithm 2 below, the effective token generation rate is obtained,
The content provider terminal, characterized in that determining the marking speed so that the marking speed is equivalent to the effective token generation speed.

A computer program combined with a computer as hardware and stored in a recording medium readable by a computer to perform the method of claim 1.

Images