ICN Working Group Xia Yong INTERNET-DRAFT China SARFT Intended Status: Informational S. Duan Expires: Sept 13, 2017 China CAICT Shu Liu China CAICT R.Huang Huawei Mar 13, 2017 the Consideration for the Application of Multi-Service Tag draft-xia-icn-multiservtag-02 Abstract This document discusses the consideration for the design of multi- service tag in the current complex network so as to optimize traffic model and improve the transmission efficiency. Status of this Memo This Internet-Draft is submitted to IETF in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/1id-abstracts.html The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html Copyright and License Notice Copyright (c) 2013 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Expires Sept 13, 2017 [Page 1] INTERNET DRAFTMar 13, 2017 Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Table of Contents 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2 Brief background . . . . . . . . . . . . . . . . . . . . . . . 3 3 The analysis of the limitation of current technologies . . . . 3 3.1 Identifying the content . . . . . . . . . . . . . . . . . . 4 3.2 Identifying the source . . . . . . . . . . . . . . . . . . . 4 4 Consideration for the typical design of multi-service tag in CDN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 4.1 the design rules of multi-service tag . . . . . . . . . . . 5 4.2 the preliminary design of multi-service tag . . . . . . . . 5 4.3 the preliminary design of the multi-service tag system . . . 7 4.4 the workflow of multi-service tag in CDN . . . . . . . . . . 8 5 Analysis of application case . . . . . . . . . . . . . . . . . 8 5.1 service perception by network . . . . . . . . . . . . . . . 8 5.2 intelligent cache . . . . . . . . . . . . . . . . . . . . . 8 5.3 content exchange between little ISPs . . . . . . . . . . . 9 6 Simple Demo . . . . . . . . . . . . . . . . . . . . . . . . . . 9 7 Security Considerations . . . . . . . . . . . . . . . . . . . . 10 8 IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 10 9 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 10 8 References . . . . . . . . . . . . . . . . . . . . . . . . . . 11 8.1 Normative References . . . . . . . . . . . . . . . . . . . 11 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 12 Expires Sept 13, 2017 [Page 2] INTERNET DRAFTMar 13, 2017 1 Introduction The document discusses the feasibility of multi-service tag in the current Internet , analyzes the limitation of the related technologies, gives the requirements of the multi-service tag. We give the typical design of multi-service tag in CDN and introduce a demo. 2 Brief background Now the network traffic presents a rapid increase trend, the popularization of network video and the diversified viewing model modes support watch video in anytime and anywhere,which also results in the increase of network traffic. The network video Apps must provide terrific Quality of experience(QoE). These trends represent a developing direction of future networks. Recognition and handling of the application traffic is a key factor for network operation. Each network application uses different protocol and is deployed by different ISP, which incompletely depends on the network operaters. The method of the recognition of traffic and applications uses the fuzzy heuristic modes which are based on the port scope and key information of the traffic and are similar with the DPI technology, but this series of technologies have some limitations. The heuristic methods can't effectively solve the problem of traffic recognition because they can't keep up with the synchronization update of application characteristics. The traffic recognition schemes based on the port scope detection face the great challenge because of enormous amount of ports which are discontinuous, especially for http traffic, the http traffic usually use 80 or 8080 port, so the content in http traffic is difficult to be identified accurately. Due to the encryption transmission of more and more traffic, these lead to the great increase of DFI/DPI calculated amount and make these two technologies be faced with invalidation. IP tunneling technology makes the operator's network more complex. So we need a new technology which can rapidly and uniquely recognize the traffic based on its characteristics without resolve the whole package. 3 The analysis of the limitation of current technologies The traffic recognition ways based on IP address pool face difficulties. Because IP address is of large amount, dynamic, proprietary or private. According to the CDN protocol (RFC 6770), the content can be transferred to different CDN and this makes it impossible to track the content among different CDN in terms of its IP address. Though the traffic recognition based on IP address is possible in some scenes, it's impossible to exactly identify every flow. Because the same port is maybe repetitively used by different Expires Sept 13, 2017 [Page 3] INTERNET DRAFTMar 13, 2017 application, the traffic recognition based on port may lead the wrong results. DFI/DPI may lose efficacy or become very complicated with the more and more encrypted traffic in order to analyze the content contained by the traffic. A traffic flow of an application will end at user terminal through different network routes and this will affect the analysis of the traffic flow. There are no unified standards for traffic recognition and analysis and it will lead to different analysis results for the same traffic flow due to the analysis ability and implementation ways. The traffic analysis will parse the payload of the packages, thus it will affect the package processing efficiency which need extra process, and the ever- increasing new protocols also affect the DFI/DPI devices efficiency. The flow tag is defined in RFC6437 and it only applies in IPv6 protocols. The flow tag changes along with the specific traffic flow and just like port. The flow tag can't identify the traffic flow independently and it must be used with source/destination IP addresses together. Because the flow tag is fixed in IPv6 header, it can identified easily, but it lacks of protect mechanism and there is no mechanism verifying its integrity. In general, the current traffic recognition ways is limited in the analysis of traffic flows, they can't provide effective feedback data, so they can't support the self-adaptive network processing capability established by the operators. 3.1 Identifying the content In order to improve the hit ratio and actively push the hot resources to the local subscribers, the cache system need a succinct way to learn the buffered contents and can judge the hot content according to the actual content information. 3.2 Identifying the source To enable flexible reverse charging, we need a third party recognizable tag of the traffic for the charging GW located between the client and server, which helps in recognition of its source and billing model, and other features to enable other cultivated transport services, e.g. QoS for selected content types for a given ICP. 4 Consideration for the typical design of multi-service tag in CDN Now the accelerating system is widely used in the Internet, such CDN, the ISP's intelligent cache and P2SP etc. These accelerating systems usually use simple URLs to speed up the access rate without regard to the specific content,bandwidth requirements, global attention etc. So there's blindness to accelerate everything using one way and there's Expires Sept 13, 2017 [Page 4] INTERNET DRAFTMar 13, 2017 no optimal acceleration aiming at the service requirements. For convenience, we use the CDN conditions to describe the application of the multi-service tag here. 4.1 the design rules of multi-service tag we need to design to a new rule to express the service properties and transmission requirements of the content. Every CDN operators can use the same algorithm to apply different accelerating strategies for the different contents so as to provide more professional and accurate accelerating service for every CP. We design a multi-service tag which can help the network better know the traffics which it carries. The multi-service tag design abandons the traditional fuzzy heuristic design idea and directly include the information about the transferred contents and the requirements for the transmission network. This kind of tag is embedded into the content URL according to some rules, it can transparently transferred in different networks and has no impact on existing services, and we call it multi-service tag. It has the following features: a) no relationship with IP address or port number; b) one-to-one correspondence to the transferred content; c) stable in a traffic flow lifecycle; d) easily obtained and handled by the network operator; e) the tag can be recognized by the network, the network can draw up a strategy and adaptively transfer the content according to the tag information; f) confidence mechanism against tamper-proofing; g) decrease the complexity of network management. 4.2 the preliminary design of multi-service tag Here we give a simple and preliminary design of multi-service tag, the scheme is not mature and may be changed along with the development of the new technologies. The format of multi-service tag is as following: Expires Sept 13, 2017 [Page 5] INTERNET DRAFTMar 13, 2017 xlables = base64( CID + content summary + type + file size + [code rate] + priority + timestamp + random number + signature ) xlables: fixed string which identifies tags and encrypts the following information using base64. CID: the identity of CP which is distributed by the tag service system in a unified way. content summary: the summary is extracted according to the file content and corresponding to the file and actually is file hash. This field can be used to identify the same cached content. type: the kinds of the transferred file, such as video, picture, document. file size: the size of the transferred file. code rate: it's specific for the video file. priority: the transferred priority level, such as normal, vip. timestamp: the time which the file is issued. random number: it provides the signature identity signature: it's produced according to the CID+content summary+type+file size+[code rate]+timestamp+random number and used to verify the validity of the tag. The tag can be embedded into the URL and the CDN can extract the tag from the URL. Through the tag service, the CDN can get the corresponding service information. Here is the example: the original URL: http://wx.cvideo.com.cn/labels/video/xvideo4k.mp4 the embedded tag URL: http://wx.cvideo.com.cn/labels/video/xvideo4k.mp4?xlabels=Y2lkPTVCRUZ CRTIxNEY3OEJBMjhERjRGRjdFNkY0RkMwRjU5Jmhhc2g9RTFFOEI3MzVFMzM3QkEyQkRCQUM1 MjQyRUUxMDc5RkImdHlwZT12aWRlbyZzaXplPTYuMzVHJmR1cmF0aW9uPTE4NjBzJnRpbWVzd GFtcD0xNDg2NjM2MDA3JnJhbmQ9ODkyMjMyMSZzaWduYXR1cmU9UTgwWjZMRFYyQkpWUVBHNl JIMFpXSU5TOU0tUEtKNUo= In fact, the length of URL is limited in CDN, so we put key information into the tag. We also design the tag related information Expires Sept 13, 2017 [Page 6] INTERNET DRAFTMar 13, 2017 corresponding to each tag which mainly includes tag service information and tag transmission information. The tag service information includes content description,drama series, VIP accelerating and copyright etc. The tag transmission information include image resolution, play software requirements, cache nodes distribution and access interests etc. The CDN extracts the tag, get the corresponding tag related information and then provides corresponding service. 4.3 the preliminary design of the multi-service tag system We design a multilevel tree structure multi-service tag system which includes tag service center and tag servers. The tag service center is the core of the system which provides storage, renewal and inquiry service for the whole network tags and the related tag information. The tag server is deployed at the networks of CP or accelerating operators which can provide distributed tag information extraction, resolution and inquiry services. |--------------------------------| | tag service center | |--------------------------------| | | | | |-------| | | | |----| |-------------| |-------------| | CP |----->| tag server |-----| | tag server |--| |----| |-------------| | |-------------| | | | | | |-------------| | | | tag server | |--------------| | |-------------| | P2SP network |---| | | |--------------| | |-----------------| | | | | CDN edge node | | | | |-----------------| | | |---------------------| | |-----------| | P2SP Speedup server | |------------------| | End User | |---------------------| | operator's cache | |-----------| |------------------| Figure 1 the multi-service tag system Expires Sept 13, 2017 [Page 7] INTERNET DRAFTMar 13, 2017 The related tag information and tag URL are usually stored at the tag server and tag service center, the normal End Users only know the normal URL, so this system has no influence on network devices and End Users. It needs the CDN/cache to do some changes. 4.4 the workflow of multi-service tag in CDN The multi-service tag system provide public tag services including tag issuing and tag recognition. The tag service can be distributed deployed so as to provide serving for the whole network accelerating services similar to DNS service. The tag issuing workflow: 1)The CP generates the tag through the tag service and embeds it to the corresponding URL; 2)The CP publishes the URL to its websites and the tag system; 3)The tag system records the tag information and updates the tag service center; 4)The tag system provides tag information inquiry service to the CDN operators through the distributed tag inquiry system. The tag recognition workflow: 1)The CDN identifies the tag field in the URL; 2)The CDN gets the basic information of the content transmission from the tag; 3)The CDN gets the tag related information through the tag service interface; 4)The CDN stores the tag's basic information and related information. 5 Analysis of application case 5.1 service perception by network The Internet video traffic becomes the major service traffic and the video traffic has high demands for the transmission network and is sensitive to the network. The 4k programs transmission needs more than 30Mbps bandwidth, it can't guarantee the QoS even though CDN is used for the transmission. The core problem is that the network isn't able to know the carried service and thus allocate resources dynamically. We can use multi-service tag to resolve the problem. Because the tag can contain the information of the carried service, the network operator can use tag to quickly identify and handle 4k program flow and demands to network to allocate resource dynamically for the 4k flow. 5.2 intelligent cache Expires Sept 13, 2017 [Page 8] INTERNET DRAFTMar 13, 2017 The cache technology is always one of the main technological means for decreasing inter-network settlement charge and enhancing QoE. The maximal challenge which the traditional cache technology faces is that the repetitive contents waste the cache resource. The core technology of the traditional cache is to obtain URL contents and store them locally by monitoring the hot program's URLs through DPI. But the URL is not stable and the same contents may have different URLs. Though we can use DPI to decode the content and acquire partial content characteristics to compare, it has major limitations at decreasing the repetitive contents and greatly increases the computation complexity, what is more, the begin of the content is often advertisement or station caption and this makes content comparison different to work well. The multi-service tag contains the attribute information of carried content which is one-to-one correspondence to the content, then the cache system can use the tag as the base of comparison so as to quickly discover the repetitive contents and raise cache efficiency. 5.3 content exchange between little ISPs In order to reduce operating costs, little ISPs are apt to interconnect each other to realize the user scale increase and resource sharing. Each little ISP has establish its own cache system or CDN node in order to decrease the inter-network settlement expense with the large ISP. The different little ISP must indicate the location of content storage and content itself, then issue the content URL. The multi-service tag not only include the information of the content and also bound together with the URL to transfer, thus resolve the problem of program content sharing after ISP interconnection. 6 Simple Demo We establish a simple demo to validate the availability of multi-service tag, irrespective of security, network complexity and other factors. In our demo, we make some simplification for the real network components and the network structure is shown in Figure 2. Test Terminal --------- content server Figure 2 Demo network structure The content server simulates the CDN edge node, its function includes:extract the tag from URL, get the tag related information, serve the user according to the related information. We store two almost same h.264 video of 4K 60P at the content server, one has normal URL and other has tag URL with VIP priority. We watch these two video files and statistics the output bandwidth. In order to watch the video normally, it needs about 30M bandwidth. For the normal URL, we see about 6M output bandwidth. For the tag URL, we see about 30M output bandwidth. Through this demo, the tag URL can work well. Expires Sept 13, 2017 [Page 9] INTERNET DRAFTMar 13, 2017 7 Security Considerations TBD. 8 IANA Considerations There is no IANA action in this document. 9 Acknowledgements TBD. Expires Sept 13, 2017 [Page 10] INTERNET DRAFTMar 13, 2017 8 References 8.1 Normative References Expires Sept 13, 2017 [Page 11] INTERNET DRAFTMar 13, 2017 Authors' Addresses Yong Xia China SARFT Email: xiayong@abs.ac.cn Shihui Duan China Academy of Telecommunication Research of MIIT Email: duanshihui@catr.cn Shu Liu China Academy of Telecommunication Research of MIIT Email: liushu@catr.cn Rachel Huang Huawei Email: rachel.huang@huawei.com Expires Sept 13, 2017 [Page 12]