WO2013155979A1 - 一种处理内容路由方法及装置 - Google Patents

一种处理内容路由方法及装置 Download PDF

Info

Publication number
WO2013155979A1
WO2013155979A1 PCT/CN2013/074421 CN2013074421W WO2013155979A1 WO 2013155979 A1 WO2013155979 A1 WO 2013155979A1 CN 2013074421 W CN2013074421 W CN 2013074421W WO 2013155979 A1 WO2013155979 A1 WO 2013155979A1
Authority
WO
WIPO (PCT)
Prior art keywords
content
routing
network
domain
routing table
Prior art date
Application number
PCT/CN2013/074421
Other languages
English (en)
French (fr)
Inventor
汪军
许欣
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2013155979A1 publication Critical patent/WO2013155979A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/50Network services
    • H04L67/56Provisioning of proxy services
    • H04L67/568Storing data temporarily at an intermediate stage, e.g. caching

Definitions

  • the present invention relates to content network related technologies, and in particular, to a method and device for processing content.
  • Cisco VSI Voice Network Index
  • Web Data data
  • file sharing file sharing
  • video-like repeatable content Web Data
  • the forwarding mode is more efficient, and the existing IP network supports an end-to-end transmission model that is inefficient under existing Internet traffic characteristics.
  • NDN uses content router networking, and uses routing protocols to exchange content routes to form a hierarchical network.
  • Some technologies do not specify the network form, but they choose flat coding.
  • Address mode such as DONA
  • routing can not be aggregated, can only be used In the form of a server cluster or a centralized directory, you cannot use a router for networking.
  • the content network in the form of a server cluster has the advantage of flexible networking and can form a cache cluster of any size.
  • the disadvantage is that the server does not transmit packets efficiently. To achieve the speed of the router, a large number of servers are required, and the energy consumption and integration are The problem, especially for the underlying network that needs to carry all the traffic on the Internet.
  • the content network in the form of a router has the advantages of fast forwarding speed and high port density.
  • the disadvantage is limited storage capacity.
  • the high-speed forwarding relies on expensive search memory and high power consumption.
  • the supported routing table has limited capacity and cannot be thinned. Granular content routing scheduling. On the other hand, the long tail effect of Internet content is obvious.
  • the embodiment of the invention provides a method for processing content routing, a network element device and a network server, so as to solve the contradiction between the content traffic forwarding efficiency and the cache hit rate.
  • the network server receives the content access information reported by the content routing device in the domain; calculates the frequency of the historically accessed content according to the content in the content access information, and adjusts the content routing table according to the frequency;
  • the adjusted content routing table is sent to all content routing devices in the domain.
  • the content is addressed as an address in an encoded form of the content identification prefix, and the network server calculates a historical access frequency of the corresponding content according to the content addressing in the content access information, and adjusts the content routing table according to the frequency, including :
  • the network server calculates a historical access frequency of the corresponding content according to the content identifier prefix;
  • the route of the specified number of contents with the highest historical access frequency is modified to point to the local cache, and the route of the specified number of contents with the lowest historical access frequency is modified to point to the corresponding router or the corresponding content source device.
  • the network server performs merge processing on the same content identifier prefix in the content routing table.
  • the above method can also have the following features:
  • the network server After receiving the routing entry of the external domain network, the network server adds the routing entry of the foreign domain network to the content routing table, and points the routing entry to the designated device in the external domain network.
  • the above method can also have the following features:
  • the network server After receiving the content access request of the content routing device, the network server queries the local cache, and if it hits, returns the locally cached content data to the content routing device; if not, queries the content routing table, according to the hit The next hop device is routed.
  • the first module is configured to receive the content access information reported by the content routing device in the domain; the second module is configured to calculate the frequency of the historically accessed content according to the content addressing in the content access information, according to the frequency Adjust the content routing table; and
  • the third module is configured to send the adjusted content routing table to all content routing devices in the domain.
  • the second module is configured to calculate a historical access frequency of the corresponding content according to the content identification prefix; modify a route of the specified number of content with the highest historical access frequency to point to the local cache, and minimize the historical access frequency.
  • the route of the specified number of contents is modified to point to the corresponding router or the corresponding content source, and is also used for combining the same content identifier prefix in the content routing table.
  • the above web server can also have the following features:
  • a fourth module configured to receive a routing entry of the external domain network, in the content routing table Add a routing entry of the foreign domain network, and point the route of the routing entry to a specified device in the external domain network.
  • the above web server can also have the following features:
  • the fifth module is configured to: after receiving the content access request of the content routing device, query the local cache, and if the hit, return the locally cached content data to the content routing device; if not, query the content routing table, Routing based on the hit next hop device.
  • Another method for processing content routing provided by the embodiment of the present invention includes:
  • the content routing device obtains a content routing table from a network center of the domain to which it belongs;
  • the content routing device After receiving the content access request, the content routing device queries the local cache, and if it hits, returns the locally cached content data to the user; if not, queries the content routing table, according to the hitting next hop device. routing.
  • the foregoing method may further have the following features: after the content routing device processes the content access request, the method further includes:
  • the content routing device samples the historical content access request information, and sends the sampled content access request information to the network center, where the content access request information includes content addressing information.
  • a first module configured to obtain a content routing table from a network center of a domain to which the content routing device belongs;
  • the second module is configured to query the local cache after receiving the content access request, and if the hit, return the locally cached content data to the user; if not, query the content routing table according to the hit next hop device Route.
  • the above content routing device may also have the following features:
  • the third module is configured to sample the historical content access request information, and send the sampled content access request information to the network center, where the content access request information includes content addressing information.
  • An embodiment of the present invention further provides a network system, including: the foregoing network server and multiple The above content routing device.
  • the embodiments of the present invention provide a method for processing content routing, a network element device, and a network server, which can resolve the contradiction between content forwarding efficiency and cache hit ratio, and achieve higher overall forwarding performance with minimum cost. Increase cache hit ratio.
  • FIG. 1 is a network architecture diagram of an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a content routing device according to an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of a network server according to an embodiment of the present invention.
  • FIG. 4 is a flowchart of a method for processing content routing according to an embodiment of the present invention.
  • FIG. 5 is a flowchart of a content router accessing a network according to an embodiment of the present invention.
  • FIG. 6 is a flowchart of content routing according to an embodiment of the present invention.
  • FIG. 7 is a flowchart of cross-domain content routing cooperation according to an embodiment of the present invention.
  • FIG. 8 is a flowchart of CRDS content routing statistics, calculation, and merging according to an embodiment of the present invention.
  • FIG. 1 shows a network architecture of an embodiment of the present invention, wherein an autonomous domain includes a content access client (eg, a PC, a mobile terminal, a tablet, etc.), a content router (CR), a content cache cluster (CCS), Content Routing Policy Decision Server (CRDS). Content routing information can be exchanged through CRDS in multiple autonomous domains.
  • a content access client eg, a PC, a mobile terminal, a tablet, etc.
  • CR content router
  • CCS content cache cluster
  • CRDS Content Routing Policy Decision Server
  • a content router (CR) is set at the edge of the network.
  • the router uses the built-in storage of the device to provide a first-level content cache.
  • the query content routing table determines the next hop to be forwarded.
  • the content router can be a dedicated network element device, or it can be a device that is modified by adding a content cache function on a mobile, fixed network access router or a common IP router.
  • CCS server-based storage cluster system
  • the capacity is set according to the overall target cache hit ratio.
  • CRDS content routing policy decision server
  • CRDS is set in the network center, which is responsible for controlling the content routing table of the content router. And collect content cache information of the content storage cluster system in the domain.
  • the content routing decision server is responsible for communicating with and exchanging routes with its domain's content routing decision server.
  • the network element device in this embodiment includes:
  • the first module is configured to obtain a content routing table from a network center of the domain to which it belongs;
  • the second module is configured to query the local cache after receiving the content access request, and if the hit, return the locally cached content data to the user; if not, query the content routing table according to the hit next hop device Route.
  • the network element device in this embodiment may further include:
  • the third module is configured to sample the historical content access request information, and send the sampled content access request information to the network center, where the content access request information includes content addressing information.
  • the next hop device includes:
  • the specified storage device, content source device, or next hop router of the network center is the first hop content cache, which caches the hottest part of the network, generally achieves a 20-40% cache hit ratio, and at the same time has a router efficient forwarding capability.
  • FIG. 3 is a schematic diagram of a network server (such as the above-mentioned CRDS) according to an embodiment of the present invention.
  • the network server of this embodiment includes: a first module, configured to receive content access reported by a content routing device in the domain.
  • the second module is configured to calculate a frequency of the historically accessed content according to the content addressing in the content access information, and adjust the content routing table according to the frequency;
  • the third module is configured to send the adjusted content routing table to all content routes in the domain. Equipment.
  • the second module is configured to calculate a historical access frequency of the corresponding content according to different content identification prefixes; and modify a route of the specified number of contents with the highest historical access frequency to point to the local cache, and the history The route of the specified number of accesses with the lowest frequency of access is modified to point to the corresponding router or corresponding content source.
  • the network server may further include:
  • the fourth module is configured to: after receiving the routing entry of the external domain network, add the routing entry of the foreign domain network to the content routing table, and point the route of the routing entry to the designated device in the external domain network.
  • the network server may further include:
  • the fifth module is configured to: after receiving the content access request of the content routing device, query the local cache, and if the hit, return the locally cached content data to the content routing device; if not, query the content routing table, Routing based on the hit next hop device.
  • FIG. 4 is a flowchart of a method for processing content routing according to an embodiment of the present invention. As shown in FIG. 4, the method includes the following steps:
  • Step 101 The content routing device obtains a content routing table from the network center of the domain to which the domain belongs; when the CR accesses the network, the CRDS contacts the CRDS, and the CRDS sends the content routing table to the CR.
  • the routing table may include:
  • the domain CCS has cached the content, or the domain name already signed (the route directly points to the local CCS);
  • Step 102 The CR processes the content access request.
  • the content routing device After receiving the content access request, the content routing device queries the local cache, and if it hits, returns the locally cached content data to the user; if not, queries the content routing table to perform routing according to the hit next hop device.
  • Step 103 After processing the content access request by the CR, the CR may sample the access information by using a certain policy. And reported to CRDS;
  • Step 104 The CRDS calculates the access frequency of the content source according to the access information reported by the CR sample, and adjusts the CR content routing table according to the access frequency.
  • the route with the highest access frequency and the original route directly pointing to the content source is adjusted to point to the local CCS.
  • CCS can prefetch the content or cache the content on demand to ensure that the hot content can hit CCS in the local domain; Point to the content source or next hop content router.
  • Step 105 The CRDS sends the adjusted content routing table to all content routing devices in the domain.
  • the above architecture and method of the embodiment of the present invention can regard the cache of the CR and the cache of the CCS as the overall cache capacity of the system, and the capacity can benefit from the distributed server cluster technology, and the capacity is large enough to enable the Zipf distributed Internet content access. Enough cache hit ratio.
  • CR is the first hop content cache, which caches the hottest part of the network. It can generally achieve 20-40% cache hit ratio and router efficient forwarding capability.
  • CR's content routing is controlled by CRDS, which can calculate content routing based on carrier content management policies and real-time statistics. It can ensure that traffic cache hit rate to CCS is high enough, and avoid unpopular content access through CCS (this part) Access basically causes Cache Miss to cause traffic to be bypassed and consumes a lot of CCS processing power.
  • the network must perceive the content address of the content access, and therefore design an inner Content access protocol
  • the content address in the content access protocol is an address in the form of a prefix encoding, for example, a content address in the form of a URL (Uniform/Universal Resource Locator) to ensure that the route can be aggregated. of.
  • the packet header is the complete path of the content, such as the content addressing form in NDN, such as content:example.com/path/xxx.ext/s_num, where s-num is for the convenience of processing large
  • the content name can also be directly addressed to the final content.
  • the content is hierarchical, which facilitates the aggregation of routes. It can be aggregated into a single instance. com/a, example. com/b. The route of example.com to reduce the total number of content routes that are delivered to the CR.
  • the CR is located at the edge, uses a router, has a high-speed forwarding line card and a local built-in cache. For the received client access request, it first looks for the local cache, and after the miss, it routes according to the content routing table. Due to the limitations of device size, power consumption, and high-speed forwarding interfaces, built-in cache capacity is often limited, and current mainstream technologies can reach terabytes (bytes) (10 12 bytes).
  • CCS is located at the center of the network. It can also be deployed in multiple geographic locations as needed.
  • server clustering technology is used in general.
  • the current mainstream technology of cache capacity can reach several terabytes/server, the total size of the cluster. Up to tens of thousands of servers and storage facilities, the total cache capacity can be on the order of PB (10 15 bytes, beat bytes) to tens of PB (10 15 bytes).
  • the long tail distribution model of Internet content also determines that the hottest content can be effectively cached with a smaller cache (CR built-in cache), and the rest of the content must have a cache size of more than PB above CCS to achieve a certain cache hit rate.
  • CRDS is located in the center of the network. It can also use the distributed deployment mode. It is responsible for controlling the content routing table of CR.
  • the main sources of information for CRDS to generate CR content routing table include:
  • the operator's predefined caching strategy such as caching for popular extra-domain websites
  • CR is a router platform
  • its routing table is often stored with special devices such as TCAM.
  • TCAM The cost and power consumption are very high, and the capacity is very limited.
  • the mainstream router TCAM capacity The quantity is in the range of tens of megabytes to hundreds of megabits. If the content is addressed in the form of a text URL, a routing entry can be up to hundreds of bytes, and the composite bit is hundreds of bits. Therefore, the content table is optimal. More than one million, so CRDS must perform route aggregation when making routing decisions, to ensure that the total number of routes is within the processing power of CR.
  • the CRDS in order to ensure that the super-cooled content access traffic is bypassed by the CCS as much as possible, the CRDS completely controls the CR routing table, and in order to adapt to the change of the content heat, the CRDS needs to perform heat statistics on the content access, but let all The access to the CRDS undoubtedly also increases the unnecessary processing overhead and the traffic detour.
  • the access statistics are reported by the CR according to the sampling rule sent by the CRDS, for example, randomly taking 2% of the Cache Miss data. , or 1% of all access information.
  • FIG. 5 is a flowchart of a content router accessing a network according to an embodiment of the present invention. The embodiment includes the following steps:
  • Step 201 CR is started, and the content routing policy decision server is assisted by manual configuration or a mechanism such as DHCP or DNS.
  • Step 202 CR and CRDS are connected, and the status and capability of the node are included, which may include information such as port status and cache capacity.
  • Step 203 The CRDS generates a preset content routing table according to the existing carrier content cache, the distribution policy, and the local content cache hotspot statistics, and directly directs the route of the pre-distribution, cache, and hot content to the local CCS, and other unpopular Content or content that is not of interest is routed to other CR or content sources;
  • Step 204 The CRDS sends the routing table and the content access sampling rule to the CR.
  • FIG. 6 is a flowchart of content routing according to an embodiment of the present invention.
  • the embodiment includes the following steps: Step 301: A client sends a content request, and the request address is a.com/b/l.rmvb;
  • Step 302 After the CR receives the request, query the local cache miss, and then query the routing table, where the routing table points to the next hop CR or content source;
  • Step 303 The CR forwards the content request to the next hop CR or the content source.
  • Step 304 The next hop CR or the content source returns the requested content
  • Step 305 The CR determines, according to a cache scheduling policy, whether to cache the content.
  • Step 306 The CR returns content to the requested client.
  • the CRDS the report information includes: content addressing, Cache Miss/Hit (cache miss/Hit); Step 308: The CRDS counts and sorts the access status of each content name according to multiple sampling reports of multiple CRs, and Further performing prefix route merging, it is found that the a.com/b path access frequency is relatively high;
  • Step 309 The CRDS informs the CCS to cache the content of the a.com/b path;
  • Step 310 The CRDS sends a routing entry to the CR, and the route of a.com/b is directed to the local CCS.
  • Step 311 The client initiates a content request of a.com/b/2.rmvb again.
  • Step 312 The CR still caches, and then queries the local routing table, and the entry already points to the ccs of the domain.
  • Steps 313-315 CCS queries the local cache, if it exists, returns directly, otherwise requests content from the content source.
  • the reason for caching 2.rmvb in CCS in this step is:
  • Steps 316-317 CCS returns the requested content to the requesting client via CR.
  • FIG. 7 is a flowchart of cross-domain content routing cooperation according to an embodiment of the present invention, including the following steps: Step 401: Domain 2 is responsible for content distribution of c.com, and local policy determines collaboration with domain 1;
  • Step 402 The CRDS of the domain 2 advertises the c.com routing entry to the CRDS of the domain 1.
  • Step 404 The client of domain 1 requests c.com related content, and sends a request to reach CR;
  • Step 405 The CR of the domain 1 queries the local cache Cache Miss, and the result points to the domain 2;
  • Steps 406-407 The CR forwards the request to the CCS of the domain 2, and the latter returns the content of the request;
  • domain 2 may also choose not to directly expose its own CCS address, but use an interworking CR to accept the request of domain 1, which does not affect the basic principles of the present invention.
  • Step 408 The CR of the domain 1 determines whether to cache the content according to the cache scheduling policy, and returns the content to the requesting client.
  • Step 409 The CR is reported to the CRDS according to the sampling policy for subsequent decision.
  • FIG. 8 is a flowchart of CRDS content routing statistics, calculation, and merging according to an embodiment of the present invention, and the specific steps are as follows:
  • Step 501 Start;
  • Step 502 The CRDS receives the content access sampling report, where the content identifier includes: the content identifier and the cache life information;
  • Step 503 Update the corresponding content access statistics by using the content identifier as a keyword.
  • Step 504 Perform merge statistics according to different length content identifier prefixes, and sort; ratio ⁇ , content www.a.com/a/l.jpg , www.a.com/a/2.jpg, www.a.com/b/1.doc Visit 4, 6, or 10 times respectively, you can merge to www.a.com/a/ to access 10 times, www .a.com/b/ Visit 10 times and further merge into www.a.com for 20 visits;
  • Step 505 Sort the prefixes whose prefix lengths are 1, 2, and 3 respectively, and take the route prefixes that are ranked first and do not overlap respectively.
  • the CRDS performs further optimization decisions according to whether the cache hits information, such as a certain content prefix. Most of the accesses are content router cache hits, indicating that the content router has cached most of the prefix content. CRDS may choose not to point this part of the content prefix to the local cache, and the local cache does not cache the data of the content prefix. Save cache capacity. Of course, CRDS can also decide not to refer to the cache to hit the information.
  • the prefix length refers to the number of segments of the content identifier.
  • the prefix length of www.a.com is 1, www.a.com/a/the length is 2, before www.a.com/a/c
  • the length of 3 ⁇ 4 is 3.
  • the prefix length is 1, 2, and 3, which is mainly due to the number of prefixes. If the number of prefixes is small in the actual network, a longer prefix can be used as the routing keyword.
  • Step 506 On the basis of the foregoing content routing prefix, the carrier policy requirement must be slowed down.
  • the stored content prefix generates a routing table.
  • the above prefix in the routing table points to CCS.
  • Other prefixes or default routes point to the next hop CR or directly to the content source.
  • Step 507 Send the generated new routing entry to the CR.
  • Step 508 End.
  • the solution of the embodiment of the present invention can solve the contradiction between content traffic forwarding efficiency and cache hit ratio, achieve higher overall forwarding performance and improve cache hit ratio with minimum cost.

Abstract

本实施例公开了一种处理内容路由的方法及装置,其中,处理内容路由的方法包括:网络服务器接收其所在域内的内容路由设备上报的内容访问信息;根据所述内容访问信息中的内容编址计算历史访问的内容的频度,根据所述频度调整内容路由表;将调整后的内容路由表下发给所述域内的所有内容路由设备。

Description

一种处理内容路由方法及装置
技术领域
本发明涉及内容网络相关技术, 特别是涉及一种处理内容路由方法及装 置。
背景技术
根据 Cisco (思科) VNI ( Visual Network Index, 视觉化网络指数 )统计 预测, 互联网流量中 90%以上的流量都是 Web Data (数据)、 文件共享、 视 频类的可重复内容, 这些流量釆用緩存、 转发的模式更为高效, 而现有的 IP 网络支持的是端到端的传输模型, 在现有互联网流量特征下是低效的。
在过去的 20年中, 已经有一些技术用于改善这些内容流量的传输, 比如 90年代的网站镜像, 90年代末出现的 Web Cache (高速緩冲存储器), 2000 年以后引入的 CDN ( Content Delivery Network, 内容分发网络 ) 均是将内容 緩存在离最终用户较近的地方, 减少报文在骨干上的重复传递, 一方面降低 了内容传输的时延、提高了用户体验,另一方面也降低了运营商骨干网开销。 但是这些技术都是互联网之上的应用层技术, 针对特定的应用、 特定的网站 进行优化, 没法改变绝大多数互联网流量端到端的传输模型。
进几年,学术界开始研究将互联网改造为以内容传输为中心的緩存 -转发 模式, 比较著名的项目包括 DONA ( Data Oriented Network Architecture, 面 向数据网络架构) 、 PSIRP ( Publish-Subsrible Internet Routing Paradigm, 发 布订阅式互联网路由模型)、 NDN ( Named Data Networking , 命名数据网络, 参见 http://www.named-data.net/ )等, 这些架构的共同特征是将主机之间通信 的基本原语从原来 IP 的连接 -发送 /接收报文的形式改为无连接的内容 /数据 读取 /订阅, 网络设备设置内容緩存, 并将解析主机发出的内容读取请求, 如 果命中本地緩存则立即返回结果, 否则继续在网络中路由。 这些架构有的明 确提出了网络实现形态, 比如 NDN釆用内容路由器组网, 釆用路由协议交 换内容路由, 形成一个层次化的网络, 有的技术没有指明网络形态, 但是其 选择了扁平化编址方式(比如 DONA ) , 这样路由比如无法汇聚, 只能釆用 服务器集群或集中的目录的形式, 无法釆用路由器进行组网。
釆用服务器集群形态的内容网络, 其优点是组网灵活, 可以组成任意规 模的緩存集群, 缺点是服务器传送报文效率不高, 要达到路由器的速度需要 大量服务器, 能耗和集成度均是问题, 尤其对于需要承载互联网所有流量的 基础网络而言更是如此。釆用路由器形态的内容网络,其优点是转发速度快, 端口密度高, 缺点是存储容量有限, 此外其高速转发所依赖的查找存储器昂 贵且功耗高, 支持的路由表容量有限, 无法进行细粒度的内容路由调度。 另 一方面互联网内容长尾效应明显, 一般认为符合 Zipf (齐普夫分布)分布, 即访问频度为, 如果釆用緩存的话, 所需緩存容量随着命中率提到呈指数上 升, 釆用路由器组网内容緩存容量受制于设备架构, 总体达到 50%的緩存命 中率即很困难。按照用户访问不重复的内容源总量为 1Ε(10Λ18)字节计算(按 照 Google (谷歌) 的统计, 2008年互联网总容量即达到 5E字节) , Zipf分 布参数 a =0.8, Cache命中率达到 50%需要接近 30ΡΒ(1Ρ=10Λ15)的緩存容量, 这远远超出了路由器设备能够提供的緩存。 发明内容
本发明实施例提供一种处理内容路由的方法及网元设备、 网络服务器, 以解决内容流量转发效率和緩存命中率之间的矛盾问题。
本发明实施例提供的一种处理内容路由的方法, 包括:
网络服务器接收其所在域内的内容路由设备上报的内容访问信息; 根据所述内容访问信息中的内容编址计算历史访问的内容的频度, 根据 所述频度调整内容路由表;
将调整后的内容路由表下发给所述域内的所有内容路由设备。
上述方法还可具有下面特点:
所述内容编址为内容标识前缀的编码形式的地址, 所述网络服务器根据 所述内容访问信息中的内容编址计算对应内容的历史访问频度, 根据所述频 度调整内容路由表, 包括:
所述网络服务器根据内容标识前缀来计算对应的内容的历史访问频度; 将历史访问频度最高的指定个数的内容的路由修改为指向本地緩存, 将 历史访问频度最低的指定个数的内容的路由修改为指向对应的路由器或对应 的内容源设备。
上述方法还可具有下面特点:
所述网络服务器对所述内容路由表中相同的内容标识前缀进行合并处 理。
上述方法还可具有下面特点: 还包括:
所述网络服务器接收到外域网络的路由表项后, 在所述内容路由表中添 加所述外域网络的路由表项, 将该路由表项的路由指向外域网络中的指定设 备。
上述方法还可具有下面特点: 还包括:
所述网络服务器接收到内容路由设备的内容访问请求后,查询本地緩存, 若命中, 则将本地緩存的内容数据返回给该内容路由设备; 若未命中, 则查 询所述内容路由表, 根据命中的下一跳设备进行路由。
本发明实施例还提供的一种网络服务器, 包括:
第一模块, 设置为接收本域内的内容路由设备上报的内容访问信息; 第二模块, 设置为根据所述内容访问信息中的内容编址计算历史访问的 内容的频度, 根据所述频度调整内容路由表; 以及
第三模块, 设置为将调整后的内容路由表下发给本域内的所有内容路由 设备。
上述网络服务器还可具有下面特点:
所述第二模块, 是设置为根据内容标识前缀来计算对应的内容的历史访 问频度;将历史访问频度最高的指定个数的内容的路由修改为指向本地緩存, 将历史访问频度最低的指定个数的内容的路由修改为指向对应的路由器或对 应的内容源 ,还用于对所述内容路由表中相同的内容标识前缀进行合并处理。
上述网络服务器还可具有下面特点: 还包括:
第四模块, 设置为接收到外域网络的路由表项后, 在所述内容路由表中 添加所述外域网络的路由表项, 将该路由表项的路由指向外域网络中的指定 设备。
上述网络服务器还可具有下面特点: 还包括:
第五模块, 设置为接收到内容路由设备的内容访问请求后, 查询本地緩 存, 若命中, 则将本地緩存的内容数据返回给该内容路由设备; 若未命中, 则查询所述内容路由表, 根据命中的下一跳设备进行路由。 本发明实施例提供的另一种处理内容路由的方法, 包括:
内容路由设备从其所属域的网络中心获取内容路由表;
所述内容路由设备接收到内容访问请求后, 查询本地緩存, 若命中, 则 将本地緩存的内容数据返回给用户; 若未命中, 则查询所述内容路由表, 根 据命中的下一跳设备进行路由。 上述方法还可具有下面特点: 所述内容路由设备处理内容访问请求后, 还包括:
所述内容路由设备对历史的内容访问请求信息进行取样, 将取样的内容 访问请求信息上才艮给所述网络中心, 所述内容访问请求信息包括内容编址信 息。
本发明实施例提供的一种内容路由设备, 包括:
第一模块, 设置为从所述内容路由设备所属域的网络中心获取内容路由 表; 以及
第二模块, 设置为接收到内容访问请求后, 查询本地緩存, 若命中, 则 将本地緩存的内容数据返回给用户; 若未命中, 则查询所述内容路由表, 根 据命中的下一跳设备进行路由。 上述内容路由设备还可具有下面特点: 还包括,
第三模块, 设置为对历史的内容访问请求信息进行取样, 将取样的内容 访问请求信息上才艮给所述网络中心, 所述内容访问请求信息包括内容编址信 息。
本发明实施例还提供了一种网络系统, 包括: 上述的网络服务器和多个 上述的内容路由设备。
综上, 本发明实施例提供一种处理内容路由的方法及网元设备、 网络服 务器, 可以解决内容流量转发效率和緩存命中率之间的矛盾, 使最少的成本 实现更高的整体转发性能并提高緩存命中率。 附图概述
图 1为本发明实施例的网络架构图;
图 2为本发明实施例的内容路由设备的示意图;
图 3为本发明实施例的网络服务器的示意图;
图 4为本发明实施例的一种处理内容路由的方法的流程图;
图 5为本发明实施例的一种内容路由器接入网络的流程图;
图 6为本发明实施例的一种内容路由的流程图;
图 7为本发明实施例的跨域内容路由协作的流程图;
图 8为本发明实施例的 CRDS内容路由统计、 计算和合并的流程图。
本发明的较佳实施方式
下文中将结合附图对本发明的实施例进行详细说明。 需要说明的是, 在 不冲突的情况下, 本申请中的实施例及实施例中的特征可以相互任意组合。
图 1示出了本发明实施例的一种网络架构, 其中一个自治域内包括内容 访问客户端 (例如, PC、 移动终端、 平板电脑等) 、 内容路由器 (CR ) 、 内容緩存集群(CCS )、 内容路由策略决策服务器(CRDS )。 多个自治域内 可以通过 CRDS交换内容路由信息。
在网络边缘设置内容路由器(CR ) , 路由器使用设备内置存储提供第一 级内容緩存, 当经过内容路由器的内容请求没有命中内置緩存时, 其查询内 容路由表决定转发的下一跳。
内容路由器可以是专门的网元设备, 也可以是移动、 固网接入路由器、 普通 IP路由器之上添加内容緩存功能改造而成的设备。 在网络中心设置基于服务器架构的存储集群系统(CCS ) , 容量根据总 体的目标緩存命中率进行设置; 同时在网络中心还设置内容路由策略决策服 务器 (CRDS ) , 它负责控制内容路由器的内容路由表, 并收集本域内内容 存储集群系统的内容緩存信息。
如需实现多域间协作, 则由内容路由决策服务器负责和它域的内容路由 决策服务器通信并交换路由。
图 2为本发明实施例的内容路由设备的示意图, 如图 2所示, 本实施例 的网元设备包括:
第一模块, 设置为从所属域的网络中心获取内容路由表;
第二模块, 设置为接收到内容访问请求后, 查询本地緩存, 若命中, 则 将本地緩存的内容数据返回给用户; 若未命中, 则查询所述内容路由表, 根 据命中的下一跳设备进行路由。 本实施例的网元设备还可以包括:
第三模块, 设置为对历史的内容访问请求信息进行取样, 将取样的内容 访问请求信息上才艮给所述网络中心, 所述内容访问请求信息包括内容编址信 息。
其中, 所述下一跳设备包括:
所述网络中心的指定存储设备、 内容源设备、 或下一跳路由器。 这样, 内容路由设备是第一跳内容緩存, 其緩存了网络中最热的部分内 容, 一般可以做到 20-40%的緩存命中率, 同时又具有路由器高效转发能力。
图 3为本发明实施例的网络服务器(如上述的 CRDS ) 的示意图, 如图 3所示, 本实施例的网络服务器包括: 第一模块, 设置为接收本域内的内容路由设备上报的内容访问信息; 第二模块, 设置为根据所述内容访问信息中的内容编址计算历史访问的 内容的频度, 根据所述频度调整内容路由表;
第三模块, 设置为将调整后的内容路由表下发给本域内的所有内容路由 设备。
其中, 所述第二模块, 是设置为根据不同的内容标识前缀来计算对应的 内容的历史访问频度; 将历史访问频度最高的指定个数的内容的路由修改为 指向本地緩存, 将历史访问频度最低的指定个数的内容的路由修改为指向对 应的路由器或对应的内容源。
在一优选实施例中, 所述网络服务器还可以包括:
第四模块, 设置为接收到外域网络的路由表项后, 在所述内容路由表中 添加所述外域网络的路由表项, 将该路由表项的路由指向外域网络中的指定 设备。
在一优选实施例中, 所述网络服务器还可以包括:
第五模块, 设置为接收到内容路由设备的内容访问请求后, 查询本地緩 存, 若命中, 则将本地緩存的内容数据返回给该内容路由设备; 若未命中, 则查询所述内容路由表, 根据命中的下一跳设备进行路由。
图 4为本发明实施例的一种处理内容路由的方法的流程图,如图 4所示, 包括下面步骤:
步骤 101、 内容路由设备从所属域的网络中心获取内容路由表; 当 CR接入到网络时主动和 CRDS联系, CRDS下发内容路由表至该 CR, 此路由表可以包括:
1 ) 、 本域 CCS已经緩存内容, 或已经签约的域名 (其路由直接指向本 地 CCS ) ;
2 ) 、 统计表明为访问热点的网站域名, 其路由指向本地 CCS。
3 ) 、 其它内容路由指向下一跳的内容路由器或内容源。
步骤 102、 CR处理内容访问请求;
内容路由设备接收到内容访问请求后, 查询本地緩存, 若命中, 则将本 地緩存的内容数据返回给用户; 若未命中, 则查询所述内容路由表, 根据命 中的下一跳设备进行路由。
步骤 103、 CR处理内容访问请求后, 可以某种策略对访问信息进行取样 并上报给 CRDS;
取样频率选择要保证流量不对 CRDS造成过高的性能压力, 也能保证统 计结果的有效性, 比如取样 x%的访问, x=l ~ 10 (即 X的取值范围为 1至 10 之间) 。
步骤 104、 CRDS根据 CR取样上报的访问信息计算内容源的访问频度, 根据访问频度调整 CR的内容路由表;
对访问频度排名靠前且原先路由直接指向内容源的路由调整为指向本域 CCS, CCS 可以预取内容或按需緩存内容, 保证热点内容能够在本域 CCS 命中; 同时将冷门内容路由直接指向内容源或下一跳内容路由器。
步骤 105 , CRDS将调整后的内容路由表下发给本域内的所有内容路由 设备。
通过本发明实施例的如上架构和方法可以将 CR的緩存和 CCS的緩存看 作系统总体的緩存容量, 此容量可以受益于分布式服务器集群技术, 容量足 够大, 从而使 Zipf分布的互联网内容访问足够的緩存命中率。
CR是第一跳内容緩存, 其緩存了网络中最热的部分内容, 一般可以做 到 20-40%的緩存命中率, 同时又具有路由器高效转发能力。
CR的内容路由受 CRDS控制, 而后者可以根据运营商内容管理策略和 实时釆样统计计算内容路由, 可以保证路由到 CCS 的流量緩存命中率足够 高,同时也避免冷门内容访问经过 CCS(此部分访问基本上会 Cache Miss(緩 存未命中) 而导致的流量迂回且占用大量 CCS处理能力。
初步计算, 在现有互联网内容模型下, 设置互联网数据总量 3%左右的 緩存 ( 1017Bytes数量级)才能达到 70%左右的緩存命中率, 也就是说 30%访 问仍然会 Cache Miss, 对于 CCS而言, 等于这 30%的流量完全穿通, 并且由 于 CCS 网络位置相对较高 (越低的话就需要设置更多的 CCS, 建设成本更 高), 在现有 IP网络之上的迂回路径也较长, 导致带宽成本增加。 因此, 本 发明中使得这原本 Cache Miss的流量可以无需经过 CCS进行迂回, 从而进 一步节省了成本。 本发明实施例中, 网络必须感知内容访问的内容地址, 因此设计一种内 容访问协议, 内容访问协议中的内容编址是一个前缀编码形式的地址, 比如 说是一个 URL ( Uniform/Universal Resource Locator, 统一资源定位符 )形式 的内容编址, 以保证其路由是可以汇聚的。
报文头部即是内容的完整路径, 比如类似于 NDN中的内容编址形式, 如 content:example.com/path/xxx.ext/s— num形式,其中 s— num是为了方便处理 大的媒体文件而设计的切片编号, 该内容名称也可以直接寻址到最终的内容 式是分层的, 方便进行路由的汇聚, 比 ¾口 example. com/a、 example. com/b可 以汇聚成为一条 example.com的路由, 以减少下发到 CR的内容路由总条目。
其中, CR位于边缘, 釆用路由器, 具有高速转发线卡和本地内置高速 緩存, 对于收到的客户访问请求, 其首先查找本地緩存, 未命中后根据内容 路由表进行路由。 由于受制于设备体积、 功耗以及高速转发接口的限制, 内 置緩存容量往往有限, 目前主流技术可以达到数 TB (字节) ( 1012字节 ) 。
CCS位于网络的中心位置, 根据需要, 其也可以釆用分布式技术部署在 多个地理位置, 但总体上釆用服务器集群技术, 緩存容量目前的主流技术可 以到数 TB/服务器, 集群总大小可达数万台服务器及存储设施, 緩存总容量 可达数 PB ( 1015字节, 拍字节)至数十 PB ( 1015字节)量级。 互联网内容 的长尾分布模型也决定了最热的内容可以用较小的緩存(CR 内置緩存) 实 现有效緩存, 其余的内容必须要 CCS这样 PB以上数量级的緩存才能达到一 定的緩存命中率。
CRDS位于网络的中心位置, 同样也可以釆用分布式部署模式, 其负责 控制 CR的内容路由表, CRDS生成 CR内容路由表的主要信息来源包括:
1、 运营商预定义的緩存策略, 比如对热门的域外网站进行緩存;
2、 运营商签约的内容网站, 运营商负责这些内容网站的内容分发; 3、 本地 CCS中已经緩存的内容;
4、 访问热度统计。
但是, 考虑到 CR是路由器平台, 其路由表往往釆用 TCAM这样的特殊 器件存储, 成本和功耗非常之高, 容量非常有限, 目前主流路由器 TCAM容 量在数十兆到数百兆比特,如果用于存储文本 URL形式的内容编址,一条路 由表项可达上百字节, 折合成比特就是数百比特, 因此, 内容表最佳是不超 过一百万条, 因此 CRDS在进行路由决策时必须进行路由汇聚, 确保路由总 量在 CR的处理能力范围内。
本发明实施例中为了保证过冷的内容访问流量尽可能不经过 CCS 进行 迂回,釆用 CRDS完全控制 CR路由表,同时为了适应内容热度的变化, CRDS 需对内容访问进行热度统计, 但是让所有访问都经过 CRDS无疑也增加了不 必要的处理开销和流量迂回, 本发明实施例釆用抽象统计的方式, 即 CR按 照 CRDS下发的抽样规则上报访问数据,比如随机取 2%的 Cache Miss数据, 或 1%的所有访问信息。
图 5为本发明实施例的一种内容路由器接入网络的流程图, 本实施例包 括以下步骤:
步骤 201 : CR启动, 通过人工配置或 DHCP、 DNS等机制辅助发现内容 路由策略决策服务器;
步骤 202: CR和 CRDS进行连接, 上^艮自己的节点状态和能力, 其中可 包括端口状态、 緩存容量等信息;
步骤 203: CRDS根据已有的运营商内容緩存、 分发策略以及本域内容 緩存热点统计信息生成了预置的内容路由表, 将预先分发、 緩存以及热点内 容的路由直接指向本域 CCS, 其它冷门内容或不关心的内容路由指向其它 CR或内容源;
步骤 204: CRDS将路由表以及内容访问取样规则下发给 CR。
图 6为本发明实施例的一种内容路由的流程图,本实施例包括以下步骤: 步骤 301 : 客户端发出内容请求, 请求地址为 a.com/b/l.rmvb;
步骤 302: CR收到请求后查询本地緩存未命中, 然后查询路由表, 路由 表指向下一跳 CR或内容源;
步骤 303: CR转发内容请求给下一跳 CR或内容源;
步骤 304: 下一跳 CR或内容源返回请求的内容;
步骤 305: 所述 CR根据緩存调度策略决定是否緩存该内容; 步骤 306: 所述 CR将内容返回给请求的客户端;
CRDS, 上报信息中包括: 内容编址、 Cache Miss/Hit (緩存未命中 /命中) ; 步骤 308: CRDS根据多个 CR的多次取样上报信息对各个内容名的访问 情况进行统计、 排序, 并进一步进行前缀路由合并, 发现 a.com/b路径访问 频度较为靠前;
步骤 309: CRDS告知 CCS对 a.com/b路径的内容进行緩存;
步骤 310: CRDS给 CR下发路由表项,将 a.com/b的路由指向本域 CCS; 步骤 311 : 客户端再次发起 a.com/b/2.rmvb的内容请求;
步骤 312: CR仍然 Cache Miss, 再查询本地路由表, 该表项已经指向本 域 ccs。
步骤 313-315: CCS查询本地緩存, 如果已有则直接返回, 否则向内容 源请求内容。 在此步骤中 CCS中緩存有 2.rmvb的原因在于:
1 ) 、 CCS釆用预取策略;
2 ) 、 已经有其它终端访问过了该内容。
步骤 316-317: CCS把请求的内容通过 CR返回给请求的客户端。
图 7为本发明实施例的跨域内容路由协作的流程图, 包括以下步骤: 步骤 401 : 域 2负责 c.com的内容分发, 并且本地策略决定和域 1进行 协作;
步骤 402: 域 2的 CRDS向域 1的 CRDS通告 c.com路由表项; 步骤 403: 域 1的 CRDS下发 c.com的路由表项给所控制的 CR, 并将其 路由表项指向域 2的 CCS;
步骤 404: 域 1的客户端请求 c.com相关内容, 发出请求到达 CR;
步骤 405: 域 1的 CR查询本地緩存 Cache Miss, 结果指向域 2;
步骤 406-407: 所述 CR向域 2的 CCS转发请求, 后者返回请求的内容; 在本步骤中,域 2也可以选择不直接暴露自己的 CCS地址, 而是釆用一 个互通的 CR接受域 1的请求, 这不影响本发明基本原理。
步骤 408: 域 1的 CR根据緩存调度策略决定是否緩存该内容, 并将内 容返回给请求的客户端;
步骤 409: 所述 CR根据取样策略上报给 CRDS, 以备后续决策。
图 8为本发明实施例的 CRDS内容路由统计、 计算和合并的流程图, 具 体步骤如下:
步骤 501 : 开始;
步骤 502: CRDS收到内容访问取样上报, 其中包括: 内容标识和緩存 是否命中等信息;
步骤 503 , 以该内容标识为关键字更新相应的内容访问统计计数; 步骤 504, 按照不同长度内容标识前缀进行合并统计, 并排序; 比^口, 内容 www.a.com/a/ l .jpg、 www.a.com/a/2.jpg、 www.a.com/b/1.doc 分别访问 4、 6、 10次,则可以合并成 www.a.com/a/访问 10次, www.a.com/b/ 访问 10次, 进一步合并成 www.a.com访问 20次;
步骤 505: 对前缀长度分别为 1、 2、 3的前缀进行排序, 分别取排名靠 前且不重叠的若干路由前缀; 同时 CRDS根据緩存是否命中信息进行进一步 的优化决策, 如某一个内容前缀中绝大部分访问都是内容路由器緩存命中, 表示内容路由器已经緩存了绝大部分该前缀内容, 则 CRDS可以选择不将此 部分内容前缀指向本地緩存, 本地緩存也不緩存此内容前缀的数据, 以节省 緩存容量。 当然, CRDS也可以决定不参考緩存是否命中的信息。
本实施例中, 前缀长度是指内容标识的分段数目, 比如 www.a.com的前 缀长度是 1 , www. a. com/a/长度是 2 , www.a.com/a/c前 ¾长度则是 3。
本实施例中, 取前缀长度为 1、 2、 3主要是出于前缀数量的考虑, 实际 网络中若前缀数量较少, 则可以取更长的前缀作为路由关键字。
步骤 506: 在上述内容路由前缀的基础上, 结合运营商策略要求必须緩 存的内容前缀, 生成路由表, 该路由表中以上前缀指向 CCS , 其它前缀或默 认路由指向下一跳 CR或直接指向内容源。
步骤 507: 将生成的新的路由表项下发给 CR;
步骤 508: 结束。
本领域普通技术人员可以理解上述方法中的全部或部分步骤可通过程序 来指令相关硬件完成, 所述程序可以存储于计算机可读存储介质中, 如只读 存储器、 磁盘或光盘等。 可选地, 上述实施例的全部或部分步骤也可以使用 一个或多个集成电路来实现。 相应地, 上述实施例中的各模块 /单元可以釆用 硬件的形式实现, 也可以釆用软件功能模块的形式实现。 本发明不限制于任 何特定形式的硬件和软件的结合。
以上仅为本发明的优选实施例, 当然, 本发明还可有其他多种实施例, 在不背离本发明精神及其实质的情况下, 熟悉本领域的技术人员当可根据本 发明作出各种相应的改变和变形, 但这些相应的改变和变形都应属于本发明 所附的权利要求的保护范围。
工业实用性 本发明实施例的方案可以解决内容流量转发效率和緩存命中率之间的矛 盾, 使最少的成本实现更高的整体转发性能并提高緩存命中率。

Claims

权 利 要 求 书
1、 一种处理内容路由的方法, 包括: 网络服务器接收其所在域内的内容路由设备上报的内容访问信息; 根据所述内容访问信息中的内容编址计算历史访问的内容的频度, 根据 所述频度调整内容路由表;
将调整后的内容路由表下发给所述域内的所有内容路由设备。
2、 如权利要求 1所述的方法, 其中:
所述内容编址为内容标识前缀的编码形式的地址, 所述网络服务器根据 所述内容访问信息中的内容编址计算对应内容的历史访问频度, 根据所述频 度调整内容路由表, 包括:
所述网络服务器根据内容标识前缀来计算对应的内容的历史访问频度; 将历史访问频度最高的指定个数的内容的路由修改为指向本地緩存, 将 历史访问频度最低的指定个数的内容的路由修改为指向对应的路由器或对应 的内容源设备。
3、 如权利要求 2所述的方法, 其中:
所述网络服务器对所述内容路由表中相同的内容标识前缀进行合并处 理。
4、 如权利要求 1所述的方法, 其还包括:
所述网络服务器接收到外域网络的路由表项后, 在所述内容路由表中添 加所述外域网络的路由表项, 将该路由表项的路由指向外域网络中的指定设 备。
5、 如权利要求 1-4任一项所述的方法, 其还包括:
所述网络服务器接收到内容路由设备的内容访问请求后,查询本地緩存, 若命中, 则将本地緩存的内容数据返回给该内容路由设备; 若未命中, 则查 询所述内容路由表, 根据命中的下一跳设备进行路由。
6、 一种网络服务器, 包括: 第一模块, 设置为接收所述网络服务器所在域内的内容路由设备上报的 内容访问信息;
第二模块, 设置为根据所述内容访问信息中的内容编址计算历史访问的 内容的频度, 根据所述频度调整内容路由表; 以及
第三模块, 设置为将调整后的内容路由表下发给所述域内的所有内容路 由设备。
7、 如权利要求 6所述的网络服务器, 其中:
所述第二模块, 是设置为根据内容标识前缀来计算对应的内容的历史访 问频度;将历史访问频度最高的指定个数的内容的路由修改为指向本地緩存, 将历史访问频度最低的指定个数的内容的路由修改为指向对应的路由器或对 应的内容源,还用于对所述内容路由表中相同的内容标识前缀进行合并处理。
8、 如权利要求 6所述的网络服务器, 其还包括:
第四模块, 设置为接收到外域网络的路由表项后, 在所述内容路由表中 添加所述外域网络的路由表项, 将该路由表项的路由指向外域网络中的指定 设备。
9、 如权利要求 6-8任一项所述的网络服务器, 其还包括:
第五模块, 设置为接收到内容路由设备的内容访问请求后, 查询本地緩 存, 若命中, 则将本地緩存的内容数据返回给该内容路由设备; 若未命中, 则查询所述内容路由表, 根据命中的下一跳设备进行路由。
10、 一种处理内容路由的方法, 包括: 内容路由设备从其所属域的网络中心获取内容路由表;
所述内容路由设备接收到内容访问请求后, 查询本地緩存, 若命中, 则 将本地緩存的内容数据返回给用户; 若未命中, 则查询所述内容路由表, 根 据命中的下一跳设备进行路由。
11、如权利要求 10所述的方法, 其中: 所述内容路由设备处理内容访问 请求后, 还包括: 所述内容路由设备对历史的内容访问请求信息进行取样, 将取样的内容 访问请求信息上才艮给所述网络中心, 所述内容访问请求信息包括内容编址信 息。
12、 一种内容路由设备, 包括:
第一模块, 设置为从所述内容路由设备所属域的网络中心获取内容路由 表; 以及
第二模块, 设置为接收到内容访问请求后, 查询本地緩存, 若命中, 则 将本地緩存的内容数据返回给用户; 若未命中, 则查询所述内容路由表, 根 据命中的下一跳设备进行路由。
13、 如权利要求 12所述的内容路由设备, 其还包括, 第三模块, 设置为对历史的内容访问请求信息进行取样, 将取样的内容 访问请求信息上才艮给所述网络中心, 所述内容访问请求信息包括内容编址信 息。
14、 一种网络系统, 包括: 如权利要求 6-9任一项所述的网络服务器和 多个如权利要求 12或 13所述的内容路由设备。
PCT/CN2013/074421 2012-04-20 2013-04-19 一种处理内容路由方法及装置 WO2013155979A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210118571.6A CN102647357B (zh) 2012-04-20 2012-04-20 一种处理内容路由方法及装置
CN201210118571.6 2012-04-20

Publications (1)

Publication Number Publication Date
WO2013155979A1 true WO2013155979A1 (zh) 2013-10-24

Family

ID=46659941

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/074421 WO2013155979A1 (zh) 2012-04-20 2013-04-19 一种处理内容路由方法及装置

Country Status (2)

Country Link
CN (1) CN102647357B (zh)
WO (1) WO2013155979A1 (zh)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102647357B (zh) * 2012-04-20 2016-04-13 中兴通讯股份有限公司 一种处理内容路由方法及装置
CN104737505B (zh) * 2013-02-05 2017-07-14 华为技术有限公司 基于缓存的路由方法及路由节点
CN103152273B (zh) * 2013-02-08 2016-04-13 杭州华三通信技术有限公司 数据缓存方法和装置
CN103152271B (zh) * 2013-04-03 2015-07-29 清华大学 一种基于内容的数据中心网络路由转发方法
CN104348722B (zh) 2013-07-31 2017-12-12 华为技术有限公司 确定内容获取路径、请求处理的方法、装置和系统
CN106063200B (zh) * 2013-12-09 2019-11-01 华为技术有限公司 一种用于内容缓存的装置、方法
CN103747046A (zh) * 2013-12-24 2014-04-23 乐视网信息技术(北京)股份有限公司 一种cdn文件管理方法、分发控制中心及内容分发系统
CN103810027B (zh) * 2014-01-24 2016-09-14 暨南大学 一种网络编码文件系统中基于数据温度的重建方法
US10348848B2 (en) * 2014-09-30 2019-07-09 Level 3 Communications, Llc Handling long-tail content in a content delivery network
WO2017096564A1 (zh) * 2015-12-09 2017-06-15 北京大学深圳研究生院 一种基于内容的集中式路由体系架构mccn
CN110035092A (zh) * 2018-01-11 2019-07-19 中国科学院声学研究所 一种icn网络中的基于lcd的隐式缓存策略
CN108924055B (zh) * 2018-08-23 2019-06-14 北京理工大学 一种基于斯坦纳树的命名数据网络多播路由方法
CN111327486B (zh) * 2018-12-17 2021-10-15 中国电信股份有限公司 流表下发方法、装置、系统及计算机可读存储介质
CN110086868B (zh) * 2019-04-25 2022-03-04 北京奇艺世纪科技有限公司 内容推送方法、装置及设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101026744A (zh) * 2007-03-30 2007-08-29 Ut斯达康通讯有限公司 分布式流媒体分发系统及流媒体内存缓冲及调度分发方法
CN101860550A (zh) * 2010-06-25 2010-10-13 复旦大学 一种基于路由器缓存的流媒体系统加速方法
CN102075562A (zh) * 2010-12-03 2011-05-25 华为技术有限公司 协作缓存的方法和装置
CN102647357A (zh) * 2012-04-20 2012-08-22 中兴通讯股份有限公司 一种处理内容路由方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101026744A (zh) * 2007-03-30 2007-08-29 Ut斯达康通讯有限公司 分布式流媒体分发系统及流媒体内存缓冲及调度分发方法
CN101860550A (zh) * 2010-06-25 2010-10-13 复旦大学 一种基于路由器缓存的流媒体系统加速方法
CN102075562A (zh) * 2010-12-03 2011-05-25 华为技术有限公司 协作缓存的方法和装置
CN102647357A (zh) * 2012-04-20 2012-08-22 中兴通讯股份有限公司 一种处理内容路由方法及装置

Also Published As

Publication number Publication date
CN102647357A (zh) 2012-08-22
CN102647357B (zh) 2016-04-13

Similar Documents

Publication Publication Date Title
WO2013155979A1 (zh) 一种处理内容路由方法及装置
US9906436B2 (en) Scalable name-based centralized content routing
EP2813060B1 (en) A method for collaborative caching for content-oriented networks
US8694675B2 (en) Generalized dual-mode data forwarding plane for information-centric network
US20160344796A1 (en) Network acceleration method, apparatus and device based on router device
Qazi et al. Optimal content caching in content-centric networks
WO2011150830A1 (zh) 获取内容的方法、节点及内容网络
WO2013029569A1 (en) A Generalized Dual-Mode Data Forwarding Plane for Information-Centric Network
Wong et al. Content routers: Fetching data on network path
WO2014000558A1 (zh) 一种cdn路由实现方法和系统
Hou et al. Bloom-filter-based request node collaboration caching for named data networking
Trajano et al. ContentSDN: A content-based transparent proxy architecture in software-defined networking
Lal et al. A cache content replacement scheme for information centric network
Zhang et al. NCPP-based caching and NUR-based resource allocation for information-centric networking
Naeem et al. Caching content on the network layer: A performance analysis of caching schemes in icn-based internet of things
Li et al. An aggregatable name-based routing for energy-efficient data sharing in big data era
Nguyen et al. Adaptive caching for beneficial content distribution in information-centric networking
Gür Energy-aware cache management at the wireless network edge for information-centric operation
WO2013113150A1 (zh) 缓存优化的方法、缓存器和缓存优化的系统
Mishra et al. A review on content centric networking and caching strategies
Zhu et al. A Popularity-Based Collaborative Caching Algorithm for Content-Centric Networking.
KR20130033252A (ko) 서비스 오버레이 네트워크에서 종단간 QoS 보장형 콘텐츠 전달 방법 및 그 시스템
Tseng et al. A novel cache scheme based on content popularity and user locality for future internet
Chen et al. Exploiting ICN for efficient content dissemination in CDNs
Ma et al. Source routing over protocol-oblivious forwarding for named data networking

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13777506

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13777506

Country of ref document: EP

Kind code of ref document: A1