JP2015076833A - Network coding device, network coding method, and network coding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a network coding device, a network coding method, and a network coding program which can enhance throughput in downloading a content.SOLUTION: A relay node 20 comprises: a cache storage unit 21 which stores a coded cache including linear combination of one or a plurality of Chunks transmitted among Chunks composing a content; a request storage unit 24 which, on receiving an Interest packet requiring any Chunk, when the cache storage unit 21 does not store a coded cache including the requested Chunk, stores the requested Chunk and data for identifying the request source; and a transfer unit 26 which transfers the Interest packet to a plurality of neighboring nodes.

Description

  The present invention relates to a network encoding device, a network encoding method, and a network encoding program in a relay node of a network.

ICN (Information-Centric Network), which is an information network configuration that realizes more efficient data transmission / spreading, has been proposed in place of the conventional Internet (see, for example, Non-Patent Document 1). In the ICN, efficient data transfer is realized by providing a relay node constituting a network with a function of caching data relayed once.
Furthermore, there is CCN (Content-Centric Network) as one of the network configurations classified as ICN. In the CCN, a node having the above-described cache function processes two types of packets called “Interest” indicating a request for content and “Data” including content data corresponding to this Interest packet.

  Further, not only the ICN but also a network relay node (for example, a router or a switch) has a mathematical operation function (encoding function) different from the copy and transfer function, thereby improving the transmission rate and throughput. In addition, a technique called network coding that realizes reduction of energy consumption has been proposed (see, for example, Non-Patent Document 2). Furthermore, it has been proposed to incorporate this network coding into the configuration of ICN (for example, see Non-Patent Documents 3 and 4). In a network encoded ICN, the encoded data is cached in the relay node.

  As a method of ICN incorporating network coding, there is a method called “CodingCache” in CCN (see, for example, Non-Patent Document 5). The content is composed of a plurality of fragments. In a normal CCN, a requesting node sends a request to one Chunk using one Interest packet. One Chunk constitutes one Data packet. On the other hand, in CCN which performed CodingCache, an Interest packet is transmitted to each Chunk as usual. However, the Data packet sent in response to the Interest packet is a linear combination of one or more Chunks including at least the designated Chunk among the plurality of Chunks constituting the content. The number of Interest packets for content is issued by the number of Chunks. The node that issued the Interest packet decodes the content by receiving a plurality of Data packets. It is known that CCN to which CodingCache has been applied is improved in network utilization efficiency compared to normal CCN.

  In CodingCache, if the node that received the Interest packet itself does not have a cache of the Data packet corresponding to the Interest packet, an adjacent node that may issue the Data packet is used by using Adaptive Forwarding. One is selected and an Interest packet is transferred to this adjacent node (for example, see Non-Patent Document 6).

B. Ahlgen, C.I. Dannewitz, C.I. Imbrenda, D.M. Kutscher, and B.C. Ohlman, “A survey of information-centric networking”, IEEE Communications Magazine, vol. 50, no. 7, pp. 26-36, 2012. V. Jacobson, D.C. K. Smetters, J.A. D. Thornton, M.M. F. Plus, N.C. H. Briggs and R.M. L. Brainard, “Networking Named Content,” Proc. CoNEXT 2009, pp. 1-12, 2009. R. Ahlswede, N.A. Cai, S.M. -Y. R. Li, and R.R. W. Yeung, “Network information flow,” IEEE Transactions on Information Theory, vol. 46, no. 4, pp. 1204-1216, 2000. M.M. Montpetit, C.I. Westphal, and D.C. Trossen, “Network coding meets information-centric networking: anarchical case for information dispersal through native network coding,” ProceedingMapping, 201. ProceedingMapping. 31-36, 2012. Q. Wu, Z. Li and G.L. Xie, “CodingCache: Multipath-aware CCN Cache with Network Coding,” Proc. ACM SIGCOMM workshop on information-centric networking 2013, pp. 41--42, Aug. 2013. C. Yi, A. Afanasyev, L.M. Wang, B.W. Zhang and L. Zhang, “Adaptive Forwarding in Named Data Networking,” ACM SIGCOMM Computer Communication Review, vol. 42, no. 3, pp. 62--67, July 2012.

  However, even with the above-described CodingCache, the utilization efficiency of the network is not sufficient, and an improvement in throughput is desired.

  An object of the present invention is to provide a network encoding device, a network encoding method, and a network encoding program that can improve the throughput in content download.

  The network encoding device according to the present invention includes a cache storage unit that stores an encoded cache based on a linear combination of one or a plurality of transmitted fragments, and a request packet that requests one of the fragments. And when there is no encoded cache including the requested fragment in the cache storage unit, a request storage unit that stores data identifying the requested fragment and the request source, and a plurality of adjacent nodes A transfer unit that transfers the request packet.

  When the network encoding apparatus receives a request packet requesting one of the fragments, and the encoded cache including the requested fragment exists in the cache storage unit, the network encoding device includes the requested fragment. An encoding unit that performs re-encoding based on a random coefficient for one or a plurality of encoding caches and transmits a data packet including the re-encoded data to the request source may be provided.

  When the encoding unit receives a data packet, any one of the encoding cache including the requested fragment included in the data packet and the content stored in the cache storage unit Re-encoding may be performed on the encoding cache including the fragment, and a data packet including the re-encoded data may be transmitted to the request source.

  The cache storage unit may store data re-encoded based on the received data packet as a new encoded cache.

  The cache storage unit may preferentially delete the linearly dependent encoded cache when the stored encoded cache exceeds a predetermined threshold.

  The cache storage unit may determine a deletion target based on a usage rate of a stored encoded cache.

  The network encoding method according to the present invention includes a step in which a relay node that transmits content stores an encoding cache based on a linear combination of one or a plurality of fragments transmitted among the fragments constituting the content; When receiving a request packet requesting any of the above, if the encoded cache including the requested fragment is not stored, storing the data identifying the requested fragment and the request source; Forwarding the request packet to an adjacent node.

  The network coding program according to the present invention stores a coding cache based on a linear combination of one or a plurality of fragments transmitted among fragments constituting the content in a relay node that transmits the content; When receiving a request packet requesting any of the above, if the encoded cache including the requested fragment is not stored, storing the data identifying the requested fragment and the request source; Forwarding the request packet to an adjacent node.

  According to the present invention, throughput in content download is improved.

It is a figure which shows the structure of the network coding system which concerns on embodiment. It is a block diagram which shows the function structure of the relay node which concerns on embodiment. It is a figure which shows the structural example of the content transmitted by the network encoding which concerns on embodiment. It is a figure which shows a function when there exists a corresponding encoding cache when the relay node which concerns on embodiment receives an Interest packet. It is a figure which shows a function when there is no corresponding encoding cache when the relay node which concerns on embodiment receives an Interest packet. When a relay node according to the embodiment receives an Interest packet, it is a diagram illustrating a function when there is no corresponding encoding cache but there is an entry requesting the same Chunk in the request storage unit. It is a figure which shows a function when the relay node which concerns on embodiment receives a Data packet. It is a figure which shows the function of the receiving node which concerns on embodiment. It is a flowchart which shows a process when the relay node which concerns on embodiment receives an Interest packet. It is a flowchart which shows a process when the relay node which concerns on embodiment receives a Data packet.

An example of an embodiment of the present invention will be described.
FIG. 1 is a diagram showing a configuration of a network coding system 1 according to the present embodiment.
In this embodiment, network coding is applied to CCN.

  Data such as contents transmitted from the transmission node 10 is received by the reception node 30 via a plurality of relay nodes 20 (network encoding devices) in the network. In the relay node 20, data encoded by an encoding process described later is stored in the cache storage unit 21. This cache storage unit 21 is referred to in response to a download request from the receiving node 30.

FIG. 2 is a block diagram illustrating a functional configuration of the relay node 20 according to the present embodiment.
The relay node 20 includes a cache storage unit 21, an encoding unit 22, a random number generation unit 23, a request storage unit 24, a transfer information storage unit 25, a transfer unit 26, and an interface 27.

The cache storage unit 21 (Cache Storage) stores the encoded cache encoded by the encoding unit 22.
The encoding unit 22 (Encoder) performs a predetermined encoding operation on the received Data packet and the encoded cache of the cache storage unit 21.
A random number generator 23 (Random Generator) generates a random number as an input to the encoding unit 22.
A request storage unit 24 (PIT; Pending Interest Table) stores request information of an Interest packet received from an adjacent node.
A transfer information storage unit 25 (FIB; Forwarding Information Base) stores a transfer destination of an Interest packet to which a Data packet may be issued.
The transfer unit 26 (Forwarder) transfers the Interest packet to the adjacent node according to the information in the transfer information storage unit 25.
The interface 27 (Face) transmits / receives an Interest packet or a Data packet to / from an adjacent node.

FIG. 3 is a diagram illustrating a configuration example of content transmitted by network coding according to the present embodiment.
The content is divided into a plurality of chunks by the transmission node 10 and transmitted. In the present embodiment, as an example, the content A will be described as including three pieces (chunks) C1 to C3.
The Data packet is a linear combination of any of these three Chunks. In the Data packet, for example, encoded information such as a content identifier and a linear combination coefficient is described in a header portion. It is assumed that all Interest packets for each Chunk are issued and a request is made for the entire content A.

Hereinafter, the function of the relay node 20 will be described in detail for each of cases 1 to 4.
<Case 1>
FIG. 4 is a diagram illustrating a function when there is a corresponding encoded cache when the relay node 20 according to the present embodiment receives an Interest packet.

The relay node 20 receives an Interest packet requesting Chunk C2 from the interface 27 (Face F1).
At this time, the cache storage unit 21 (Cache Storage) stores one or more encoded caches that are linear combinations of Chunks including the requested Chunk C2.

The encoding unit 22 (Encoder) performs re-encoding using a plurality of encoding caches extracted from the cache storage unit 21 and the random coefficient generated by the random number generation unit 23 (Random Generator) as input of the encoding cache. Generate a random linear combination.
The encoding unit 22 sends back this random linear combination as a Data packet corresponding to the received Interest packet.
Note that the encoding unit 22 may perform re-encoding including an encoding cache including a Chunk (C1 or C3) related to the same content other than the requested Chunk.

<Case 2>
FIG. 5 is a diagram illustrating a function when there is no corresponding encoded cache when the relay node 20 according to the present embodiment receives an Interest packet.

The relay node 20 receives an Interest packet requesting Chunk C2 from the interface 27 (Face F1).
At this time, the cache storage unit 21 does not store the encoded cache including the requested Chunk C2.

  Further, the relay node 20 has not received an Interest packet for Chunk C2 in the past, and there is no unprocessed entry for Chunk C2 in the request storage unit 24 (PIT; Pending Interest Table). Therefore, the relay node 20 additionally stores in the request storage unit 24 information indicating which Chunk the Internet packet is received from which interface 27 (request from Face F1 to Chunk C2).

Subsequently, the transfer unit 26 acquires, from the transfer information storage unit 25 (FIB: Forwarding Information Base), a list of interfaces 27 for adjacent nodes or networks to which a Data packet including Chunk C2 may be issued. Then, the transfer unit 26 transfers the Interest packet from these interfaces 27 (for example, Faces F2 and F3) to the adjacent node.
Note that all or a predetermined number of the plurality of interfaces 27 acquired from the transfer information storage unit 25 may be selected at random as the transfer destination.

<Case 3>
FIG. 6 is a diagram illustrating a function when there is no corresponding encoding cache when the relay node 20 according to the present embodiment receives an Interest packet, but there is an entry requesting the same Chunk in the request storage unit. is there.

The relay node 20 receives an Interest packet requesting Chunk C2 from the interface 27 (Face F1).
The relay node 20 has received an Interest packet for Chunk C2 in the past, and there is an unprocessed entry (for example, a request from Face F2) regarding Chunk C2 in the request storage unit 24 (PIT; Pending Interest Table). .

  The relay node 20 additionally stores, in the request storage unit 24, information indicating which Chunk received the Internet packet from which interface 27 (request from Face F1 to Chunk C2). In this case, since the Interest packet for Chunk C2 has already been transmitted, no further transfer processing is performed.

<Case 4>
FIG. 7 is a diagram illustrating a function when the relay node 20 according to the present embodiment receives a Data packet.

The relay node 20 receives a Data packet including Chunk C2 from the interface 27 (Face F3).
At this time, it is assumed that the cache storage unit 21 (Cache Storage) stores another encoded cache that is a linear combination of any one of the Chunks (C1 or C3) constituting the content A.

  The relay node 20 checks whether there is an entry corresponding to the received Data packet in the request storage unit 24 (PIT). Then, the encoding unit 22 (Encoder) generates re-encoded encoding caches corresponding to the number of entries, and the interface 27 (Face F1 and F2) requesting the Interest packet described in each entry. To transmit a Data packet.

Specifically, the encoding unit 22 includes a Data packet received from Face F3, an encoding cache related to the content A extracted from the cache storage unit 21, and a random coefficient generated by a random number generation unit 23 (Random Generator). As an input, a new linear combination of encoding caches corresponding to the number of entries is generated.
When there is no target encoding cache in the cache storage unit 21, re-encoding for the received Data packet can be omitted.

  Further, the relay node 20 additionally stores the re-encoded encoding cache in the cache storage unit 21. At this time, if the specified threshold value is exceeded, the cache storage unit 21 overwrites the old encoded cache. Alternatively, the cache storage unit 21 may preferentially delete any encoding cache having a linear dependency relationship or an encoding cache having a low utilization rate.

FIG. 8 is a diagram illustrating functions of the receiving node 30 according to the present embodiment.
The reception node 30 includes a cache storage unit 31 (Cache Storage), a decryption unit 32 (Decoder), and an interface 33 (Face F4).

The receiving node 30 issues an Interest packet for all Chunks (C1 to C3) constituting the content A, receives a Data packet as a response thereto, and stores it in the cache storage unit 31.
In this example, the receiving node 30 receives a data packet related to Chunk C3 from the interface 27 (Face F4) following Chunks C1 and C2.

  The decoding unit 32 extracts the Chunk encoded cache constituting the content A from the cache storage unit 31, and decodes it by solving the linear combination to obtain the content A.

  FIG. 9 is a flowchart showing processing when the relay node 20 according to the present embodiment receives an Interest packet.

  In step S <b> 1, the relay node 20 receives an Interest packet from the adjacent node via the interface 27.

  In step S <b> 2, the relay node 20 determines whether or not the encoded cache including the chunk requested by the received Interest packet is stored in the cache storage unit 21. If this determination is YES, the process proceeds to step S3, and if the determination is NO, the process proceeds to step S5.

  In step S3, the relay node 20 re-encodes the encoding cache including the requested Chunk by the encoding unit 22, and generates a new encoding cache.

  In step S4, the relay node 20 transmits the generated new encoded cache to the request source as a Data packet.

  In step S <b> 5, the relay node 20 determines whether another request related to the requested Chunk exists in the request storage unit 24 (PIT). If this determination is YES, the process proceeds to step S6, and if the determination is NO, the process proceeds to step S7.

  In step S6, the relay node 20 adds an entry corresponding to the received Interest packet to the request storage unit 24 (PIT), and ends the process.

  In step S7, the relay node 20 adds an entry corresponding to the received Interest packet to the request storage unit 24 (PIT).

  In step S8, the relay node 20 extracts information on the interface 27 to be transferred from the transfer information storage unit 25 (FIB), and determines a plurality of transfer destinations.

  In step S9, the relay node 20 transfers the Interest packet to the plurality of interfaces 27 determined in step S8.

  FIG. 10 is a flowchart showing processing when the relay node 20 according to the present embodiment receives a Data packet.

  In step S <b> 11, the relay node 20 receives a Data packet from the adjacent node via the interface 27.

  In step S <b> 12, the relay node 20 extracts from the cache storage unit 21 an encoded cache related to the same content as the Chunk included in the received Data packet.

In step S13, if there is an entry requesting a chunk included in the received data packet in the request storage unit 24 (PIT), the relay node 20 executes random re-encoding for the number of entries. Create a new encoding cache.
Note that re-encoding may be omitted when there is no entry in the request storage unit 24 (PIT) and when there is no encoded cache related to the cache storage unit 21.

  In step S14, the relay node 20 transmits a Data packet including a new encoding cache to the request source of the Chunk specified by the entry of the request storage unit 24 (PIT).

  In step S15, the relay node 20 deletes the entry related to the Data packet transmitted in step S14 from the request storage unit 24 (PIT).

  In step S16, the relay node 20 additionally stores the encoded cache generated in step S13 in the cache storage unit 21.

  As described above, according to the present embodiment, the relay node 20 as the network encoding device includes the cache storage unit 21. When the relay node 20 does not store the requested Chunk when receiving the Interest packet, the relay node 20 stores the requested Chunk and the request information indicating the request source, and then the Data packet including this Chunk. The Interest packet is transferred to a plurality of adjacent nodes that can be expected to be issued.

That is, in the conventional technique (CodingCache), only one possible neighboring node is randomly selected. In the present embodiment, all of the possible neighboring nodes or a plurality of arbitrarily selected neighboring nodes are selected. In contrast, the Interest packet is transferred.
Therefore, according to the method for transferring an Interest packet according to the present embodiment, compared to a case in which a corresponding Interest packet is issued using only a single path and a corresponding Data packet is received only from this single path, the relay node 20 can acquire more kinds of linear combinations, and can transmit and spread content efficiently. As a result, even if the cache capacity is equivalent to the CCN, it is possible to expect an improvement in throughput in content download, and more reliable content decoding is ensured.
Further, the spread of the encoding cache forms a download system that is resistant to network failures.

When the relay node 20 returns a Data packet to the Interest packet, the relay node 20 performs re-encoding based on a random coefficient for a plurality of encoding caches including the requested Chunk and transmits the result to the request source. .
In the conventional technique (CodingCache), since this re-encoding process is not performed, nodes that receive an Interest packet for the same Chunk always return the same Data packet. On the other hand, in the encoding method of the present embodiment, the relay node 20 performs re-encoding with a random coefficient. Therefore, even if a plurality of Interest packets are issued, a plurality of Data issued from adjacent nodes are issued. Packets are linearly independent with high probability. Therefore, since the probability that the encoded caches collected at the time of decoding are linearly independent from each other is improved, it is possible to expect an improvement in the network utilization efficiency with respect to sending of the entire content.

Similarly, when the relay node 20 receives the Data packet for the Interest packet, the relay node 20 performs re-encoding including the encoding cache stored for the same content, and transmits the encoded packet to the request source.
As a result, since the issued Data packet becomes linearly independent with a high probability, an improvement in network utilization efficiency can be expected for sending the entire content.

At this time, the relay node 20 stores the data re-encoded based on the received Data packet in the cache storage unit 21 as a new encoded cache.
In the conventional technique (CodingCache), a single linear combination is generated, cached, and transferred. However, in the encoding and caching method of the present embodiment, the relay node 20 generates a plurality of different linear combinations, Keep updating the cache. As a result, since the issued Data packet becomes linearly independent with a high probability, an improvement in network utilization efficiency can be expected for sending the entire content.

Further, the relay node 20 includes means for optimizing the cache storage unit 21, and when a predetermined threshold is provided for the cache capacity, priority is given to a linearly dependent encoded cache or an encoded cache having a relatively low utilization rate. Can be deleted.
Therefore, it can be expected that the utilization efficiency of the cache storage unit 21 is improved and the linear independence of the encoded cache spread in the network is increased.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. Further, the effects described in the present embodiment are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the present embodiment.

  In the present embodiment, the content is configured by a plurality of chunks. However, in order to suppress a decrease in transmission efficiency due to an increase in the number of chunks, for example, a plurality of contents are provided according to the content capacity. It is also possible to perform network coding by linear combination of a plurality of Chunks for each group.

  Moreover, although this embodiment demonstrated the case where network coding was applied to CCN, this invention is applicable to the data center service or cloud service in various networks.

  The network encoding method according to the present embodiment is realized by software. When realized by software, a program constituting the software is installed in an information processing apparatus (computer). Also, these programs may be recorded on a removable medium such as a CD-ROM and distributed to the user, or may be distributed by being downloaded to the user's computer via a network.

DESCRIPTION OF SYMBOLS 1 Network coding system 10 Transmission node 20 Relay node (network coding apparatus)
21 cache storage unit 22 encoding unit 23 random number generation unit 24 request storage unit 25 transfer information storage unit 26 transfer unit 27 interface 30 receiving node 31 cache storage unit 32 decoding unit 33 interface

Claims (8)

A cache storage unit that stores an encoded cache based on a linear combination of one or a plurality of transmitted fragments among the fragments constituting the content;
When a request packet requesting one of the fragments is received, if an encoded cache including the requested fragment does not exist in the cache storage unit, data for identifying the requested fragment and a request source is stored. A request storage unit,
And a transfer unit that transfers the request packet to a plurality of adjacent nodes.   When an encoded cache including the requested fragment is present in the cache storage unit when a request packet requesting any of the fragments is received, one or more encoded caches including the requested fragment The network encoding device according to claim 1, further comprising: an encoding unit that performs re-encoding based on a random coefficient and transmits a data packet including the re-encoded data to the request source.   When the encoding unit receives a data packet, any one of the encoding cache including the requested fragment included in the data packet and the content stored in the cache storage unit The network encoding device according to claim 2, wherein re-encoding is performed on an encoding cache including a fragment, and a data packet including the re-encoded data is transmitted to a request source.   The network encoding device according to claim 3, wherein the cache storage unit stores data re-encoded based on the received data packet as a new encoding cache.   5. The network encoding device according to claim 1, wherein the cache storage unit preferentially deletes a linearly dependent encoded cache when the stored encoded cache exceeds a predetermined threshold.   The network encoding device according to claim 5, wherein the cache storage unit determines a deletion target based on a usage rate of a stored encoded cache. A relay node that transmits content
Storing a coding cache by linear combination of one or more transmitted fragments among the fragments constituting the content;
Storing the data identifying the requested fragment and the requester if the encoded cache including the requested fragment is not stored when a request packet requesting any of the fragments is received; ,
Transferring the request packet to a plurality of adjacent nodes. To relay nodes that transmit content,
Storing a coding cache by linear combination of one or more transmitted fragments among the fragments constituting the content;
Storing the data identifying the requested fragment and the requester if the encoded cache including the requested fragment is not stored when a request packet requesting any of the fragments is received; ,
And a step of transferring the request packet to a plurality of adjacent nodes.

Images