CN107835129B - Content center network edge node potential energy enhanced routing method - Google Patents
Content center network edge node potential energy enhanced routing method Download PDFInfo
- Publication number
- CN107835129B CN107835129B CN201711002678.3A CN201711002678A CN107835129B CN 107835129 B CN107835129 B CN 107835129B CN 201711002678 A CN201711002678 A CN 201711002678A CN 107835129 B CN107835129 B CN 107835129B
- Authority
- CN
- China
- Prior art keywords
- content
- node
- potential energy
- cache
- publisher
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000005381 potential energy Methods 0.000 title claims abstract description 82
- 238000000034 method Methods 0.000 title claims abstract description 18
- 230000002708 enhancing effect Effects 0.000 claims abstract description 6
- 230000007246 mechanism Effects 0.000 claims description 38
- 235000008694 Humulus lupulus Nutrition 0.000 claims description 22
- 238000011144 upstream manufacturing Methods 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 10
- 239000000126 substance Substances 0.000 claims description 9
- 230000008859 change Effects 0.000 claims description 4
- 230000007774 longterm Effects 0.000 claims description 3
- 238000004088 simulation Methods 0.000 description 11
- 230000004044 response Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 230000001965 increasing effect Effects 0.000 description 5
- 230000009467 reduction Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010835 comparative analysis Methods 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- WLNBMPZUVDTASE-HXIISURNSA-N (2r,3r,4s,5r)-2-amino-3,4,5,6-tetrahydroxyhexanal;sulfuric acid Chemical compound [O-]S([O-])(=O)=O.O=C[C@H]([NH3+])[C@@H](O)[C@H](O)[C@H](O)CO.O=C[C@H]([NH3+])[C@@H](O)[C@H](O)[C@H](O)CO WLNBMPZUVDTASE-HXIISURNSA-N 0.000 description 1
- 101000741965 Homo sapiens Inactive tyrosine-protein kinase PRAG1 Proteins 0.000 description 1
- 102100038659 Inactive tyrosine-protein kinase PRAG1 Human genes 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/14—Routing performance; Theoretical aspects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
- H04L67/63—Routing a service request depending on the request content or context
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to a content center network edge node potential energy enhanced routing method, and belongs to the field of internet. Aiming at the problem that a content center network cannot sense cache content to cause low routing efficiency, an ENPER based on potential energy is provided, a potential energy model of a node is established, and an interest packet is attracted to a nearby cache node to respond by enhancing the potential energy of an edge node, so that the purposes of reducing the forwarding time of the interest packet and improving the cache hit rate are achieved; in addition, ENPER counts and predicts the popularity of the content through the edge node, and distinguishes potential energy announcement ranges of the content with different popularity by combining the size of the network, so as to achieve the purpose of reducing network overhead. The invention effectively reduces the load of the publisher server and the expense of cache announcement, and the average request delay of the content is reduced by 43 percent compared with Best-routing.
Description
Technical Field
The invention belongs to the field of internet, and relates to a potential energy enhanced routing method for edge nodes of a content center network.
Background
According to the predicted report of cisco vni, the traffic generated by the video-like application in 2021 will account for about 82% of the total network traffic. The need for mass-replication propagation of video content has led to the popularity and commercialization of Content Distribution Networks (CDNs) and peer-to-peer networks (eyedropper, P2P). Both of the two methods improve the speed of the user for accessing the content, but the CDN forwards the user request to an edge server of the network in a DNS redirection manner, and the storage location of the content is limited; P2P generates a tracker for each content that is rendered less effective. Since both cannot get rid of the end-to-end forwarding mode based on the IP address, security accidents such as DDoS attacks are caused. Therefore, researchers have proposed a new future network architecture, Information Centric Networking (ICN), that fundamentally solves the conflict between the current connection-oriented internet model and the rapidly increasing traffic demand. The ICN is identified by the name of the content rather than the assigned IP address, so that the request made by the user only needs to be concerned with the content itself, not with the location where the content is stored. Typical ICN architectures are DONA, puruit, CCN, COMET, PSIRP, where CCN (contentcentricnetwork) is considered one of the most promising approaches.
The CCN architecture adopts a naming mode similar to URL, provides service from an end to content, supports the function of expanding routing nodes, enables the router to have a traditional forwarding function and certain storage capacity, and aims to reduce the transmission time of a data packet in a network in a mode of 'storage and bandwidth exchange' and realize a distributed cache network which is more flexible than a CDN. However, in the conventional CCN routing mechanism, the advantages of such "distributed cache" are not well utilized, because the conventional routing mechanism can only implement routing of interest packets to publishers, and a large amount of nearby caches outside the path cannot be used by sensing, which results in a large amount of waste of bandwidth resources. Therefore, an efficient and reliable cache-aware routing mechanism needs to be designed to fully exploit the caching advantages of CCN.
For the mechanism of cache perception of CCN, the main problem to be solved can be summarized into two points: 1) how to discover cached content; 2) how to forward the interest package towards the nearest one of the content sources.
The existing cache-aware routing mechanisms can be divided into two categories, one is that a requester actively issues a message to detect the position of cache content: the existing literature provides a routing mechanism (NCE) for neighbor cache detection, and the scheme calculates the shortest path by using a distributed ant colony algorithm, so that the sensing of local cache can be realized. However, this solution does not explicitly indicate the depth of probing, which may cause excessive overhead when the network range is increased. The other type is that the cached content information is issued to the periphery by the cache node, the requester carries out comprehensive judgment after passively receiving the content information, and then an optimal path is selected: the literature proposes a Potential energy based routing mechanism (CATT), which constructs a Permanent Potential Field (PPF) for a stable publisher node, and is implemented in a flooding notification manner similar to the conventional CCN; a Variable Potential Field (VPF) is constructed for a variable cache node, potential energy of content of a neighbor node is announced by adopting a fixed hop count, and an interest packet determines a forwarding port of a next hop according to the received minimum potential energy so as to acquire the latest content. However, the scheme does not distinguish the server performance of the cache node and the publisher node, so that the interest packet does not select the nearest cache node route, and the request time delay of the content is large; meanwhile, the CATT advertises all cached contents to surrounding nodes by the same hop count, which causes huge bandwidth overhead.
Since the main function of the CCN router is still to forward packets quickly, the limitation of the router caching function needs to be considered. In addition, in most real scenes, the routers in the core and the middle of the network do not generate content requests, and the interest packets are all from users near the edge of the network, so that the interest packets are attracted to the nearby cache nodes as much as possible by using the cache at the edge of the network, and the response speed of the content is improved.
Disclosure of Invention
In view of this, the present invention provides a routing method for enhancing potential energy of an edge node in a content centric network, and provides a routing mechanism for enhancing potential energy of an edge node.
In order to achieve the purpose, the invention provides the following technical scheme:
the method for enhancing the routing of the content center network edge node potential energy comprises the following steps:
s1: the interest packet reaches the cache node;
s2: the server sends a corresponding data packet according to the received interest packet and returns the data packet along the original path;
s3: establishing an Edge Node enhanced Routing (ENPER) Potential energy model;
s4: and diffusing the potential energy by utilizing a cache node content announcement mechanism.
Further, the S1 specifically includes:
when an interest packet reaches a certain node, a Content storage table (CS) is inquired according to the routing characteristics of a Content Central Network (CCN); if the relevant items are matched, returning the data packet directly; if the relevant entries are not matched, inquiring a Pending Interest Table (PIT);
if the PIT inquires that the prior interest packet requests the content, adding a request port into a PIT entry and waiting for the return of a data packet; if the corresponding matching item is not inquired in the PIT, adding a request entry in the PIT, and searching a Forwarding interface with the minimum potential value in a Forwarding Information Base (FIB) for Forwarding by the interest packet; if the query is not available in the FIB at the moment, discarding the interest packet;
further, the S2 specifically includes:
after receiving the interest packet, the publisher server sends a corresponding data packet and returns the data packet along the original path; checking the PIT every time when one hop passes, and if a plurality of ports exist in the PIT, copying a data packet and sending the data packet to a plurality of requesters; the data packets are then stored in the CS and a node potential field within the autonomous domain is established.
Further, the S3 specifically includes:
abstracting the network into an undirected graph, and when a new content k is generated, dividing all nodes in the network into content publisher nodes npCache node ncAnd node n without content cacheiObtaining superposed potential energy; s represents a node set of the cache content k:caching potential energy parameters of nodes Wherein the content of the first and second substances,the value range of (1) is between 0 and 1; l is the size of the load of the router, when the request quantity of the interest packet is too large and the router can not meet the request, L is reduced,reducing the parameter value;for any number of cache node to publisher hops,caching nodes for edgesTo publisher server npThe number of hops in between; the closer to the cache at the edge of the network,the closer the value is to 1, the larger the absolute value of the potential energy is; the closer to the cache of the publisher is,the closer the value is to 0, the smaller the absolute value of the potential energy.
Further, the publisher node npThe potential energy calculation method comprises the following steps:
setting a publisher server as a long-term stable negative point charge to form a full-network potential field in the autonomous domain; when the content is not changed, the full network potential field keeps a stable state; any node n in the networkiIs received by the publisher npHas an attractive force ofAnd calculating according to the hop count, the time delay or the link bandwidth between the two:
wherein the content of the first and second substances,the quality of the content is produced for the publisher,is a node niTo the publisher node npA minimum number of hops in between; n is a symbol when n ispThe calculation formula of the potential energy isWhen is niThe calculation formula of the potential energy is
Node n further away from the publisher as the number of hops increasesiThe smaller the potential energy attractive force is, i.e.The smaller.
Further, the cache node ncThe potential energy calculation method comprises the following steps:
setting α as cache node ncContent mass ratio coefficient of (2):
wherein the content of the first and second substances,the quality of the content is produced for the publisher,is a node niTo the publisher npα is constant between 0.1 and 0.3.
Further, the calculation method of the superimposed potential energy comprises the following steps:
wherein the content of the first and second substances,the quality of the content is produced for the publisher,is a node niTo the publisher node npα is constant between 0.1 and 0.3.
Further, the S4 specifically includes:
the content popularity of a CCN network is calculated according to the number of requests for content, assuming that a k content is at a requesting node niThe number of interest packet requests received in the last certain time period T isThe popularity of the content is defined as:
wherein K represents the content type of the cache node,indicating the arrival of n within a T periodiTotal number of requests; and correcting the popularity by adopting a simple prediction mechanism, wherein sigma is a prediction weight:
PT+1(k)=σPT(k)+(1-σ)PT-1(k)
when k content is requested for the first time, the edge node collects the total number of requests of the downstream, and notifies the popularity of the k content of the upstream node while sending an interest packet upstream; after the content returns and the potential field is established, the edge nodes continue to count in the time period of the period T, and inform the upstream cache nodes of the change of the popularity of the content, so that the real-time performance of the popularity of the upstream is maintained, and the uniformity of the same content announcement range is ensured.
Further, the announcement scope satisfies:
when P is presentT+1(k) If the number of the advertised range hops is 0, the node does not advertise the content if the number of the advertised range hops is less than H1;
when Hm is less than PT+1(k) If the number of hops in the notification range is m < Hm +1, the notification is carried out on the hop range of m around, wherein m is the threshold number of the potential energy notification normal hops, and m is 2,3,4 …; n is the threshold number of maximum potential energy announcement hop count;
when P is presentT+1(k) When Hn is higher, the hop count of the notification range is n, namely the notification is carried out on the nodes in the n-hop range around;
wherein H1, H2, … and Hn are jump thresholds, H1< H2< … < Hn; hm is the threshold and k is the name of the requested content.
The invention has the beneficial effects that:
1) on the basis of potential energy concept proposed by the existing literature, a routing mechanism for enhancing the potential energy of an edge node is proposed aiming at the problem that an interest packet in a CCN comes from a network edge but cannot obtain a quick response at the edge node.
2) The method provides a mechanism for predicting the popularity of the content at an edge node and carrying out different hop counts on the potential energy value of the cached content by combining the size of the network topological range.
Drawings
In order to make the object, technical scheme and beneficial effect of the invention more clear, the invention provides the following drawings for explanation:
FIG. 1 illustrates a forwarding process of an NPER;
FIG. 2 is a potential energy overlay of a cache node;
FIG. 3 is a potential field diagram of a topology and cache nodes of a network;
FIG. 4 is a graph of average request delay versus simulation time;
FIG. 5 is a graph of average request delay versus Zipf;
FIG. 6 is a graph of publisher server load reduction rate versus Zipf variation;
fig. 7 shows the trend of the cache advertisement overhead with Zipf.
Detailed Description
Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
In the CCN routing mechanism, when a publisher generates a new content, the publisher advertises the nodes of the whole network in a flooding manner, and an interest packet looks up an optimal path to the publisher based on a Forwarding Information Base (FIB). For each router, only the next-hop port to the content publisher node is included in the FIB, but not to the nearby cache node. The average request delay of the traditional CCN routing mechanism is large because of the inability to sense the large amount of cache resources present in the network. On the basis of potential energy concepts proposed by scholars such as Suyong Eum and the like, the invention redesigns a cache-sensible edge node enhanced routing mechanism. Under the model, a publisher or a cache object is taken as negative point charge, and an interest packet, namely probe charge with positive charge, is forwarded along the direction with the fastest gradient decline and the smallest potential energy.
Fig. 1 shows a forwarding process of an interest packet when the interest packet passes through a cache node. When an interest packet reaches a certain node, firstly inquiring a Content storage table (CS) according to the routing characteristics of the CCN, and directly returning a data packet if the Content storage table is matched with a related entry; querying a Pending Interest Table (PIT) if no matching item exists, (1) if the prior Interest packet is queried in the PIT to request the content, adding a request port into a PIT entry, and waiting for the return of a data packet; and (3) if the PIT does not have a corresponding matching item, performing step (2) to add a request entry in the PIT, and then (3) the interest packet looks up the forwarding interface with the minimum potential value in the FIB for forwarding. If not queried in the FIB at this time, (4) the packet of interest will be discarded.
And after receiving the interest packet, the publisher server sends a corresponding data packet and returns the data packet along the original path. After checking the PIT, copying the data packet to a plurality of requesters (6) if a plurality of ports exist in the PIT, storing the data packet in the CS, and establishing a node potential field in the autonomous domain.
Establishing an ENPER potential energy model
Firstly, abstract the network into undirected graph, when generating a new content k, all nodes in the network can be divided into three types, namely a content publisher node npCaching node ncAnd node n without content cacheiAnd S represents a node set of the cache content k:
1.1 potential energy of publisher node
The publisher server is a source of content generation, has a large output rate, high processing performance and long-time storage capacity, and is set as a long-term stable negative point charge to form a full potential field in the autonomous domain. The full potential field, once established, will remain stable unless the content changes, and need not be updated as often. Any node n in the networkiIs received by the publisher npHas an attractive force ofThe value is calculated according to the distance between the two, and the distance can be hop count, time delay, link bandwidth and the like.
Wherein the content of the first and second substances,the negative sign of equation (1) ensures that interest packets are routed towards the lowest point of the potential field, for the quality of the content produced by the publisher, the size being related to factors such as the cache capacity of the server, the processing speed, etc.Is defined as a node niTo npThe node n which is farther from the publisher as the hop count increasesiThe smaller the potential energy attractive force is, i.e.The smaller.
1.2 caching potential energy of nodes
Although the route forwarding node in the CCN network has caching capability, compared with the publisher, the main function of the router is to forward the data packets at line speed, and the cache capacity and processing performance of the router are far inferior to those of the publisher server providing the content for the local domaincContent mass ratio coefficient of (2):
for example, when the total number of the contents in the network is N1000 and Zipf 1.0, the cumulative probability of the requests for the first 129 contents reaches 0.8, and the relationship between the other Zipf distribution indexes and the number of the contents when the cumulative probability reaches 0.8 is shown in table 1.
TABLE 1 relationship between Zipf distribution index and the number of contents when the request accumulation reaches 0.8
1.3 superposition of potential energy
Since the data packet will be stored in the pass-through node in the return, when there are multiple cache contents and contents generated by the publisher in the network at the same time, the total potential field will be presented in a linear superposition manner, as follows:
it can be inferred from the above equation that when there are multiple caches, the potential energy near the publisher is greater than that of the replica node near the edge by overlapping each other, as shown in fig. 2 (a). However, in most practical scenarios, the interest packets come from users near the edge network, and the routers in the core and middle of the network do not generate requests; if interest packets from the edge are attracted by the network core and forwarded to the publisher server, cache contents closer to the edge will be missed, resulting in an increase in user request delay. Therefore, the potential energy value of the cache at the edge of the reinforcing network is considered, as shown in the right part of fig. 2 (b). The reasons are two reasons: 1) the interest packets from the edge are directly hit on the edge node, so that the request time delay can be reduced; 2) the requests of the edge cache nodes for collecting the same interest packets for multiple times in a centralized manner increase the residence time of the content, improve the availability of the target cache content and further increase the cache hit rate.
Fig. 3 shows a potential field diagram of the topology of a content-centric network. When a new content k is generated in the network, the publisher first advertises its flood to all routers within the autonomous domain and establishes a routing entry in the router's FIB to the publisher. Assuming that the PC1 first sends out an interest packet requesting k contents, since no node in the network has the cached contents at this time, the routing node will calculate the shortest path according to the FIB and forward the interest packet to the publisher, and the responding data packet will return to the PC1 hop by hop along the request path of the previous interest packet and establish a potential field. The cache here adopts a CCN default scheme CCE (cacheeventing every), that is, the content k is stored in all nodes passed by on the return path,when the PC2 issues the same interest package request k, it will be paired on RAAndand judging the potential energy value of the issued cache content. Setting up the publisherThe content quality proportion coefficient of the cache node is α ═ 0.1, and when the network does not adopt the edge node potential energy strengthening mechanism, the cache node receives the content quality proportion coefficient on RAThe potential energy released isReceive fromThe potential energy released isDue to the fact thatThe next hop of the interest package is RB, towards the one with less potential energyForwarding, not in the nearestAnd (4) hitting.
According to the analysis, in order to ensure that the interest packet can be attracted by the edge cache potential energy to route towards the nearest cache, the potential energy parameter of the cache node is providedAnd under the condition of considering the node load, improving the potential value of the edge cache.
Is in the range of 0 to 1. Wherein, L is the size of the load of the router, when the request quantity of the interest packet is too large and the router can not meet the request, L is reduced,the parameter value is decreased.For any number of cache node to publisher hops,caching nodes for edgesTo publisher server npThe number of hops in between. The closer to the cache at the edge of the network,the closer the value is to 1, the larger the absolute value of the potential energy is; the closer to the cache of the publisher is,the closer the value is to 0, the smaller the absolute value of the potential energy. The same publisher as in FIG. 3The content quality scaling factor α of the cache node is 0.1, the router can meet the request requirement, that is, when L is 1, the potential energy of the edge node is enhanced, and the potential energy value of the interest packet sent to the RA after being mixed and received is equal to that after being received Since at this timeAfter the interest packages are compared, the interest packages face the direction of lower potential energy and fewer hopsAnd (6) forwarding.
2. Cache node content advertisement mechanism
After the model of the node potential field is established, if an advertisement mechanism is lacked to diffuse out potential energy, the potential field-based routing mechanism is not different from the traditional CCN routing mechanism: the interest packet will hit randomly on the forwarding path, and the adjacent cache node is missed. Therefore an announcement mechanism needs to be added to achieve attraction of the potential field. However, if all the cached content is advertised to the whole network, not only a large amount of overhead of the network is caused, but also when the cached content is replaced according to different algorithms, NACK advertisement needs to be issued to surrounding nodes to delete corresponding FIB entries, which also occupies a large amount of bandwidth resources. Therefore, a simple and efficient popularity determination mechanism and an advertisement range mechanism adapting to the network topology are needed to implement potential energy based routing.
2.1 computation and prediction of popularity of advertised content
In the existing literature, content is divided into three classes according to content requests in a power law distribution feature: popular, ordinary and cold, according to which the cached content is distinguished and announced, but the scheme is established by knowing the number of requests and the overall popularity of all the content in advance, and cannot be realized under the real network condition. In addition, the requested number of interest packets also has "convergence", and when a node receives a large number of the same interest packets, the downstream request port is recorded and added in the PIT, and then only one interest packet is sent upstream, so a manner similar to the statistics of the number of interest packets received by an upstream node or all nodes in a statistical domain proposed in the literature is also not preferable. According to the above analysis, the popularity value of the content can be counted and calculated only at the edge node. The edge-enhanced potential energy scheme provided by the invention can attract interest packets with the same request to an edge cache node as much as possible under the condition of considering node load, so that more accurate content popularity is obtained.
The content popularity of a CCN network is calculated according to the number of requests for content, assuming that a k content is at a requesting node niThe number of interest packet requests received in the last certain time period T isThen the popularity of the content is defined as:
wherein K represents the content type of the cache node,indicating the arrival of n within a T periodiTotal number of requests. Since a period of time elapses from the hit of the interest packet to the issuance of the notification to the surrounding nodes, and the popularity may change in this period of time, the popularity needs to be corrected by a simple prediction mechanism, where σ is the prediction weight:
PT+1(k)=σPT(k)+(1-σ)PT-1(k) (6)
when k content is requested for the first time, the edge node will collect the total number of requests downstream and inform the upstream node of the popularity of k content while sending an interest packet upstream. After the content returns and the potential field is established, the edge nodes continue to count in the period T, and inform the upstream cache nodes of the change of the content popularity, so that the real-time performance of the upstream popularity is maintained, and the uniformity of the same content informing range is ensured.
2.2 setting of Notification Range of cache nodes
As can be seen from the above section, when the expected popularity is larger, the number of representative requests is larger, the stability of the cached content is higher, and the wide-range notification thereof can improve the availability of the content. The maximum range n-hop setting of the advertising node needs to depend on a specific network topology and size, and needs to meet the following requirements: 1) the advertised scope is limited within a domain; 2) the control flow generated by the duplicate notification is limited to a certain extent; 3) the choice of n should be less than the number of hops between the caching node to the publisher.
Table 2 cache advertisement scope
Therefore, the invention sets the hop threshold value according to the size of the network to realize the cache announcement of the content with different popularity, and the threshold value is H1, H2, …, Hn (H1)<H2<…<Hn). When P is presentT+1(k)<H1, the node does not advertise the content, when PT+1(k)>When Hn, it announces to nodes in n-hop range around, when Hm<PT+1(k)<Hm +1, the device informs the m hop range around, wherein 0<k<n。
3. Experimental setup
In this embodiment, an open-source simulation platform ndnSim is used to implement the Routing mechanism, and the CATT and the main-stream Routing mechanism Best-Routing on the simulation platform are compared. And the ndnSim is a simulation tool based on NS-3, and an NDN protocol stack is added on the platform, so that a routing mechanism of the CCN network can be realized. The total number of nodes in the experiment is 30, and the request arrival obeys Poisson distribution. The distribution of the requests of the users is furthermore adjusted according to the exponential distribution of Zipf. The Zipf index represents the concentration of requested contents, and the larger the index is, the more similar the requested contents are, and the smaller the index is, the more dispersed the requested contents are. The simulation time was 180s, and the other parameters are shown in table 3:
table 3 experimental parameter settings
3.1 Performance evaluation index
In order to objectively reflect the actual effects of different routing mechanisms and the influence of different parameters on the routing mechanisms, the following performance evaluation indexes are defined in the embodiment.
1) Average request latency
The request delay of a single content refers to the time from the beginning of sending the interest packet to the whole process of receiving the data packet, and the average request delay refers to the average value of all the request delay of the content in the range with the period T (set as 20s here). The average request delay reflects the average time from the request sending to the response sending of the user, and the smaller the value, the shorter the waiting time is represented, and the better the user experience is.
2) Publisher Server Load Reduction Rate (Load Reduction Ratio)
Where S _ counts represents the number of responses of the publisher server, and R _ counts represents the total number of requests of the user. The reduction rate of the server load represents that the load of the publisher is reduced due to the response of the distributed caches in the network, and the higher the index is, the more obvious the effect of the caches in the network is.
3) Cache advertisement Message Overhead (Overhead of cache Notification Message)
The cache announcement overhead defines the product of the length of each cache announcement message and the transmission distance in unit time, and sums the number of the announcement contents, wherein the size mainly depends on the length of the message, the announcement number of the message and the hop count. The larger the value, the more overhead the cache advertisement takes, and the more bandwidth it occupies.
3.2 simulation results and analysis
Setting the simulation condition that the cache announcement of the CATT is 2 hops, the maximum announcement range of ENPER is 3 hops, and Zipf is 1. In the initial stage of simulation, because all routing nodes in the network have no storage, interest packets of different routing mechanisms must reach the publisher server to acquire contents, the request time delay in the initial stage is equal and larger, the hop count of the acquired contents is reduced along with the gradual increase of the cached contents in the network, the average time delay is gradually reduced, and finally the contents tend to be stable. Through comparative analysis, the average request time delay of the three routing mechanisms is Best-routing, CATT and ENPER from large to small. The specific reasons are as follows: in the FIB of Best-routing, only the shortest path to the publisher exists, and the content of the cache node outside the path cannot be sensed, so that most interest packets need to pass through the whole network to reach the publisher, the occupied link resources are the most, and the average request time is the longest; the CATT adopts a potential energy-based cache perception routing mechanism, and compared with Best-routing, the CATT can enable an interest packet to be routed towards a cache node, and the average request delay is reduced; the ENPER attracts the interest packet to the nearest cache node hit by increasing the potential value of the edge node, the hop count is minimum, and the occupied resources are minimum.
According to the simulation result of fig. 4, it can be analyzed that the data fluctuation is large at the initial stage of the simulation, and the data is the simulation preheating time, so that the stable data between 100 seconds and 180 seconds is taken for the subsequent comparison to be analyzed. Because the index distribution of the Zipf has certain difference in different network scenes, the difference of the three routes is compared by changing the distribution parameters (0.7-1.1) of the Zipf. As shown in fig. 5, as the Zipf parameter increases, the average delay of three requests is continuously reduced, because the smaller the Zipf distribution index is, the more discrete the request content is, the diversified content request will cause the limited storage space to be replaced with high frequency, and the cache utilization rate is low; with the increase of the Zipf distribution index, the locality and the concentration of the request content are continuously enhanced, the content stored by the CS is stable, the interest packet can hit the content in the cache node, and the average request delay is continuously reduced. The comparative analysis can show that when Zipf is 1, average request latency of ENPER is reduced by about 43% compared to Best-routing and 17% compared to CATT.
Fig. 6 analyzes the impact of the Zipf distribution index on publisher server load. A higher publisher load reduction rate indicates that the interest package may have more content requested on the caching nodes in the network. The best performance of the three comparisons is ENPER. When Zipf is 1, the ENPER routing mechanism can reduce the load of the publisher server by 83% because the interest packet can get the requested content at the edge node without reaching the publisher by changing the potential of the node. As the Zipf index increases, the rate of increase of the value corresponding to the ordinate is slowed down because the cache mechanism adopted is LRU, that is, when the router cache is full, the least recently requested content is eliminated. When the requests are gradually concentrated and the buffer capacity size is kept constant, the growth trend is slow.
Fig. 7 analyzes the relationship between the cache advertisement overhead and the Zipf exponential distribution. Compared with the prior art, the advertising cost of ENPER is lower, because the CATT can diffuse the contents of all cache nodes to the surrounding nodes in a fixed hop count mode within the cycle time, and does not distinguish the non-popular contents with few request times, and the blind advertising mode wastes bandwidth resources. Because the potential energy of the edge cache node is increased by the ENPER, the interest packets can be hit on the edge node through the attraction of the potential energy, the number of the interest packets can be collected in the edge node in a centralized manner, the request distribution of a user can be counted better, and the PIT filtering effect existing in an upstream node is reduced. In addition, ENPER sets the announcement range according to the prediction value of popularity, and a large amount of non-popular content cannot send announcement messages to surrounding nodes, so that the availability of the content is improved, and the announcement overhead is reduced.
In order to realize the nearby response of the request content and improve the utilization rate of cache resources, potential energy models are constructed for cache nodes and publisher nodes of a content center network, and a routing mechanism ENPER with enhanced edge node potential energy is provided on the basis. By changing the potential value of the node close to the edge, the attraction of the edge cache content is increased, the interest packet is attracted to the nearby node to quickly hit and respond, and the average request delay of the content and the load of a publisher server are reduced; meanwhile, the predicted popularity of the content is calculated at the edge node, the range announcement of the content with less quantity and high popularity is expanded, the announcement of the content with more quantity and low popularity is reduced or not sent, and the consumption of the announcement message on the network bandwidth is reduced.
Finally, it is noted that the above-mentioned preferred embodiments illustrate rather than limit the invention, and that, although the invention has been described in detail with reference to the above-mentioned preferred embodiments, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the invention as defined by the appended claims.
Claims (6)
1. The method for enhancing routing potential energy of the edge node of the content center network is characterized by comprising the following steps: the method comprises the following steps:
s1: the interest packet reaches the cache node;
s2: the server sends a corresponding data packet according to the received interest packet and returns the data packet along the original path;
s3: establishing an ENPER potential energy model of an edge node enhanced routing mechanism;
s4: diffusing the potential energy by utilizing a cache node content announcement mechanism;
the S1 specifically includes:
when the interest packet reaches a certain node, inquiring a content storage table CS according to the routing characteristics of the content center network CCN; if the relevant items are matched, returning the data packet directly; if the relevant entries are not matched, inquiring a pending interest table PIT;
if the PIT inquires that the prior interest packet requests the content, adding a request port into a PIT entry and waiting for the return of a data packet; if the corresponding matching item is not inquired in the PIT, adding a request entry in the PIT, and searching a forwarding interface with the minimum potential energy value in a forwarding information table (FIB) for forwarding the interest packet; if the query is not available in the FIB at the moment, discarding the interest packet;
the S2 specifically includes:
after receiving the interest packet, the publisher server sends a corresponding data packet and returns the data packet along the original path; checking the PIT every time when one hop passes, and if a plurality of ports exist in the PIT, copying a data packet and sending the data packet to a plurality of requesters; then storing the data packet in the CS, and establishing a node potential field in the autonomous domain;
the S3 specifically includes:
abstracting the network into an undirected graph, and when a new content k is generated, dividing all nodes in the network into content publisher nodes npCache node ncAnd node n without content cacheiObtaining superposed potential energy; s represents a node set of the cache content k:caching potential energy parameters of nodes Wherein the content of the first and second substances,the value range of (1) is between 0 and 1; l is the size of the load of the router, when the request quantity of the interest packet is too large and the router can not meet the request, L is reduced,reducing the parameter value;for any number of cache node to publisher hops,caching nodes for edgesTo the publisher node npThe number of hops in between; the closer to the cache at the edge of the network,the closer the value is to 1, the larger the absolute value of the potential energy is; the closer to the cache of the publisher is,the closer the value is to 0, the smaller the absolute value of the potential energy.
2. The content centric network edge node potential energy enhanced routing recited in claim 1The method is characterized in that: the publisher node npThe potential energy calculation method comprises the following steps:
setting a publisher server as a long-term stable negative point charge to form a full-network potential field in the autonomous domain; when the content is not changed, the full network potential field keeps a stable state; any node n in the networkiBy the publisher node npHas an attractive force of And calculating according to the hop count, the time delay or the link bandwidth between the two:
wherein the content of the first and second substances,the quality of the content is produced for the publisher,is a node niTo the publisher node npA minimum number of hops in between; n is a symbol when n ispThe calculation formula of the potential energy isWhen is niThe calculation formula of the potential energy is
3. The content centric network edge node potential energy enhanced routing method of claim 1, wherein: the cache node ncThe potential energy calculation method comprises the following steps:
setting α as cache node ncContent mass ratio coefficient of (2):
4. The content centric network edge node potential energy enhanced routing method of claim 1, wherein: the calculation method of the superimposed potential energy comprises the following steps:
5. The content centric network edge node potential energy enhanced routing method of claim 1, wherein: the S4 specifically includes:
the content popularity of a CCN network is calculated according to the number of requests for content, assuming that a k content is at a requesting node niThe number of interest packet requests received in the last certain time period T isThe popularity of the content is defined as:
wherein K represents the content type of the cache node,indicating the arrival of n within a T periodiTotal number of requests; and correcting the popularity by adopting a simple prediction mechanism, wherein sigma is a prediction weight:
PT+1(k)=σPT(k)+(1-σ)PT-1(k)
when k content is requested for the first time, the edge node collects the total number of requests of the downstream, and notifies the popularity of the k content of the upstream node while sending an interest packet upstream; after the content returns and the potential field is established, the edge nodes continue to count in the time period of the period T, and inform the upstream cache nodes of the change of the popularity of the content, so that the real-time performance of the popularity of the upstream is maintained, and the uniformity of the same content announcement range is ensured.
6. The content centric network edge node potential energy enhanced routing method of claim 5, wherein: the announcement scope satisfies:
when P is presentT+1(k) If the number of the advertised range hops is 0, the node does not advertise the content if the number of the advertised range hops is less than H1;
when Hm is less than PT+1(k) If the number of hops in the notification range is m < Hm +1, the notification is carried out on the hop range of m around, wherein m is the threshold number of the potential energy notification normal hops, and m is 2,3,4 …; n is the threshold number of maximum potential energy announcement hop count;
when P is presentT+1(k) When Hn is higher, the hop count of the notification range is n, namely the notification is carried out on the nodes in the n-hop range around;
wherein H1, H2, … and Hn are jump thresholds, H1< H2< … < Hn; hm is the threshold and k is the name of the requested content.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711002678.3A CN107835129B (en) | 2017-10-24 | 2017-10-24 | Content center network edge node potential energy enhanced routing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711002678.3A CN107835129B (en) | 2017-10-24 | 2017-10-24 | Content center network edge node potential energy enhanced routing method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107835129A CN107835129A (en) | 2018-03-23 |
CN107835129B true CN107835129B (en) | 2020-06-02 |
Family
ID=61649044
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711002678.3A Expired - Fee Related CN107835129B (en) | 2017-10-24 | 2017-10-24 | Content center network edge node potential energy enhanced routing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107835129B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108449608B (en) * | 2018-04-02 | 2020-12-29 | 西南交通大学 | Block downloading method corresponding to double-layer cache architecture |
CN110213315B (en) * | 2018-05-30 | 2022-01-07 | 腾讯科技(深圳)有限公司 | Path determining method and device |
CN108881943B (en) * | 2018-06-07 | 2020-08-14 | 南京邮电大学 | Node cache replacement method based on elimination index in P2P streaming media on-demand system |
CN108769252B (en) * | 2018-06-21 | 2020-11-27 | 河南科技大学 | ICN network pre-caching method based on request content relevance |
CN109302348B (en) * | 2018-09-03 | 2021-08-31 | 中国联合网络通信集团有限公司 | Data processing method based on CNN network and router |
CN111404972B (en) * | 2019-01-02 | 2023-04-07 | 中国移动通信有限公司研究院 | Data processing method, device, equipment and storage medium |
CN109981460B (en) * | 2019-03-19 | 2020-10-02 | 北京邮电大学 | Service-oriented converged network, calculation and storage integrated method and device |
CN110365801B (en) * | 2019-08-26 | 2021-12-17 | 陕西师范大学 | Partition-based cooperative caching method in information center network |
CN111556514B (en) * | 2020-04-14 | 2021-09-21 | 北京航空航天大学 | Decentralized mobile edge computing resource discovery and selection method and system |
CN112579301B (en) * | 2020-12-28 | 2022-07-01 | 北京理工大学 | NDN-based data transmission method for P2P distributed machine learning training system |
CN113382053B (en) * | 2021-04-23 | 2022-07-19 | 河南科技大学 | Content active pushing method based on node semi-local centrality and content popularity |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101710895A (en) * | 2009-11-30 | 2010-05-19 | 中国人民解放军信息工程大学 | Method for obtaining multiple next hop routing, device and router |
CN105657054A (en) * | 2016-03-04 | 2016-06-08 | 重庆大学 | Content center network caching method based on K means algorithm |
CN106028057A (en) * | 2016-05-05 | 2016-10-12 | 北京邮电大学 | Caching method for adaptive streaming content of scalable coding in mobile CCN (Content-Centric Network) |
WO2016170559A1 (en) * | 2015-04-23 | 2016-10-27 | Nec Corporation | Method and system for bandwidth reservation for multiple tree based traffic engineering in content centric networks |
CN106657181A (en) * | 2015-10-30 | 2017-05-10 | 中国科学院声学研究所 | CCN (Content center network) based data push method |
CN106850436A (en) * | 2017-01-10 | 2017-06-13 | 中国矿业大学 | Mine hybrid wireless mesh network routing protocols based on virtual potential energy field |
-
2017
- 2017-10-24 CN CN201711002678.3A patent/CN107835129B/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101710895A (en) * | 2009-11-30 | 2010-05-19 | 中国人民解放军信息工程大学 | Method for obtaining multiple next hop routing, device and router |
WO2016170559A1 (en) * | 2015-04-23 | 2016-10-27 | Nec Corporation | Method and system for bandwidth reservation for multiple tree based traffic engineering in content centric networks |
CN106657181A (en) * | 2015-10-30 | 2017-05-10 | 中国科学院声学研究所 | CCN (Content center network) based data push method |
CN105657054A (en) * | 2016-03-04 | 2016-06-08 | 重庆大学 | Content center network caching method based on K means algorithm |
CN106028057A (en) * | 2016-05-05 | 2016-10-12 | 北京邮电大学 | Caching method for adaptive streaming content of scalable coding in mobile CCN (Content-Centric Network) |
CN106850436A (en) * | 2017-01-10 | 2017-06-13 | 中国矿业大学 | Mine hybrid wireless mesh network routing protocols based on virtual potential energy field |
Non-Patent Citations (3)
Title |
---|
Replication management and cache-aware routing in information-centric networks;Vasilis Sourlas et al;《2014 IEEE Network Operations and Management Symposium (NOMS)》;20140509;全文 * |
内容中心网络中基于差异化缓存通告的混合路由机制;葛国栋等;《电子与信息学报》;20150315;全文 * |
内容中心网络兴趣包泛洪攻击理论分析模型研究;赵炯;《中国优秀硕士学位论文全文数据库》;20170115;全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN107835129A (en) | 2018-03-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107835129B (en) | Content center network edge node potential energy enhanced routing method | |
Wong et al. | Neighborhood search and admission control in cooperative caching networks | |
An et al. | An in-network caching scheme based on energy efficiency for content-centric networks | |
CN105656788B (en) | CCN content buffering method based on popularity statistics | |
Wu et al. | Multi-path selection and congestion control for ndn: An online learning approach | |
Yamamoto | A survey of caching networks in content oriented networks | |
Kim et al. | Cache capacity-aware content centric networking under flash crowds | |
Zhang et al. | A distributed in-network caching scheme for P2P-like content chunk delivery | |
Chen et al. | BRR-CVR: A collaborative caching strategy for information-centric wireless sensor networks | |
Hu et al. | Opportunistic on-path caching for named data networking | |
Gupta et al. | ICN based efficient content caching scheme for vehicular networks | |
WO2020160007A1 (en) | Semantics and deviation aware content request and multi-factored in-network content caching | |
KR20140099834A (en) | A method and system for adaptive content discovery for distributed shared caching system | |
Nguyen et al. | Adaptive caching for beneficial content distribution in information-centric networking | |
Alduayji et al. | PF-EdgeCache: Popularity and freshness aware edge caching scheme for NDN/IoT networks | |
Naz et al. | Multi-attribute caching: towards efficient cache management in content-centric networks | |
Zhu et al. | A Popularity-Based Collaborative Caching Algorithm for Content-Centric Networking. | |
Li et al. | Effective caching schemes for minimizing inter-ISP traffic in named data networking | |
CN107135271B (en) | Energy-efficient content-centric network caching method | |
Li et al. | A smart routing scheme for named data networks | |
Gu et al. | Node Value and Content Popularity‐Based Caching Strategy for Massive VANETs | |
Alahmadi | A New Efficient Cache Replacement Strategy for Named Data Networking | |
Van et al. | MS-CCN: Multi-source content centric networking | |
Chen et al. | Gain-aware caching scheme based on popularity monitoring in information-centric networking | |
Wei et al. | A novel dynamic adaptive video streaming solution in content-centric mobile network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200602 |