Keyword-Based Breadcrumbs: A Scalable Keyword-Based Search Feature in Breadcrumbs-Based Content-Oriented Network Kévin Pognart, Yosuke Tanigawa, Hideki Tode Dept. of Computer Science and Intelligent Systems Osaka Prefecture University Osaka, Japan {pognart@com., tanigawa@, tode@}cs.osakafu-u.ac.jp Abstract—The Internet shows limited performances for users’ needs especially on content sharing and video streaming. Content-Oriented Networks (CONs) are efficient approaches for such uses. They abandon the location-based routing of the Internet (IP routing) for a content identifier-based routing. In CONs, users must know the exact content identifier to request it. To give users an easier use of CONs, we propose the Keyword-based Breadcrumbs (KBC), a scalable keyword- based retrieval function for CONs based on Breadcrumbs (BC). Our work focuses on BC because of its simplicity, scalability, particularity and because it can be deployed in today’s network, even partially. KBC uses stored information about contents in routers to retrieve several possible answers to a keyword-based request while keeping original behavior of BC for content identifier-based request. We present in this paper the working scheme of KBC, some KBC request managing rules to collect answers, and simulations results to show its performances. Keywords-Breadcrumbs; Content-Oriented Network; search; keyword; cache. I. INTRODUCTION The current Internet has a host-to-host architecture for allowing an easy communication between two machines. But in the recent years, people use the Internet not for direct communication but principally for sharing contents with a lot of people and viewing video streams. As a result, the current Internet has limited performances. Nowadays, some systems such as peer-to-peer (BitTorrent) can improve content sharing performances by coordinating several users by the contents they have. Inspired by peer-to-peer systems, Content-Oriented Network (CON) is an alternative to the current Internet with some special features. Content identifier is used instead of location identifier for routing messages. Also, content identifiers are unique. The main characteristic of CON is that contents can be copied and cached in the network while keeping the same content identifier. It allows answering a request by one of the content copies located over the network. As for CON, several routing methods have been proposed to realize it [1][2]. In our work, we particularly focus on Breadcrumbs (BC) [3][11] due to its attractive features described in Section II. This BC-based CON has simple content caching, location and routing systems. In BC, we assume that users and possibly routers have a content cache. Routers have also a BC table used to route requests. When content passes through a router, a BC entry is created in its BC table to indicate the direction of the cached content. If the content goes through a node having a content cache, the content is cached. Requests are firstly sent to a server to download contents by using IP routing. When a request arrives at a router where a BC entry for the same content identifier exists, the request is redirected to follow the direction shown in the BC entry. Each next node will redirect the request according to the direction in BC entries until finding the content in a content cache. If an issue occurs during the redirection, the BC entries are invalidated and the request is forwarded again to the server by IP routing. To perform the routing, content identifiers must be unique. This uniqueness makes the requests difficult from a user’s point of view. This problem also exists in the current Internet with URLs, and it leads to the need to use web search engines. Current web search engines are not an efficient solution because they use location and they cannot use cached content information. Hence, we propose Keyword-based Breadcrumbs (KBC). We extensively designed the BC framework to complement it with a keyword-based search feature while keeping the way of working of BC and its advantages. Also, we have implemented different KBC request behaviors to retrieve answers and we compare their performances. In this paper, we present CONs and some unknown content search features. Then, we describe principle, specifics, and settings of KBC, and we compare KBC and the unknown content search features presented. We continue by evaluating KBC performances with some simulation scenario. To conclude, we summarize the main points about KBC and we talk about our future work. II. RELATED WORK A. Related CON schemes To create a CON, several schemes have been proposed. The Data oriented Network Architecture (DONA) [5], the Network of Information (NetInf) [6], the Publish-Subscribe Internet Routing Paradigm (PSIRP) [7], and the Content- Centric Networking (CCN) [4] are the main approaches. In DONA, sources publish contents into the network and their information is spread to the nodes called resolution handlers. A request goes to a resolution handler to be routed to the 20 Copyright (c) IARIA, 2014. ISBN: 978-1-61208-377-3 AFIN 2014 : The Sixth International Conference on Advances in Future Internet