The Survey, Research Challenges, and Opportunities in ICN

Supratik Banerjee

NIIT University, Neemrana, India

Tapan Naskar

NIIT University, Neemrana, India

Sanjay Kumar Biswash

NIIT University, Neemrana, India

This chapter throws light on the limitations of current Internet architecture (TCP/IP) and highlighting the importance of Information-Centric Networking (ICN), as it is used to overcome the limitations of legacy networks architecture. The primary attributes of ICN are the unique naming scheme, in-network caching, routing and forwarding. It is the most usable and best choice for the future Internet architecture. Therefore, the significance of ICN and its association with the Internet of Things (loT) will be thoroughly discussed in this chapter. We will emphasize the research challenges, real-time development and implementation of ICN-loT architecture. Finally, we propose the implementation and challenges of edge computing and cloud computing w.r.t. ICN.

Introduction

The current Internet architecture (TCP/IP) has been designed for host-to-host communication as file transfer was the main design goal of TCP/IP. Thus TCP/IP is ineffectual to adapt to the current challenges i.e., momentous increase in information distribution over the network and variability in Internet traffic. These challenges propelled the idea of Information-Centric Networking' (ICN) which proposes the exchange of content[1] [2] rather than communication between hosts and network devices [150].

To improve the availability and performance of the current Internet infrastructure, technologies like Content Distribution Networks and P2P content distribution applications evolved. Both of these technologies are promoting a communication model of accessing data by name, regardless of origin server location. However, their implementation as an overlay adversely impacts the overall efficiency. Moreover, they rely on proprietary distribution technologies. Thus the identification of named information is dependent on the type of distribution channel. Hence, technological innovations to support the identification of named information which is independent of the channel and the source became the need of the hour.

ICN is an approach to evolve the Internet infrastructure to directly support this use by introducing uniquely named content as a core Internet principle. ICN supports data independence from location, application, storage, and means of transportation, enabling in-network caching and replication. Better scalability and improved efficiency are the expected benefits of ICN [67]. ICN also offers a promising solution to problems like security and network congestion. Its chief design goals like unique naming, multicast communication and in-network caching minimize the response latency and server load.

Moreover, ICN provides promising future for loT networks because of its unique features like naming mechanism which supports the naming of devices, data, and services for loT, better security and privacy, in-network caching and in-network processing which benefits the resource constraint loT devices. And because of decoupling between the sender and receiver in ICN, data can be transmitted in case of intermittent network connectivity of mobile loT devices [312]. M2M communication in loT has been already tested on ICN [258]. Moreover, loT applications like VANET suffer from a frequent disconnection problem which is taken care of by the in-network caching facility in ICN. Thus ICN is envisioned to be an appropriate architecture for modern network schema such as loT networks [285]. Even though ICN's future is quite promising, its large- scale deployment is facing several challenges like generating name before the content creation in live streaming, updating, and versioning named data objects, accessibility management, privacy protection of requester, data origin authentication, access control, mobility management and implementation of in-network caching.

The rest of the chapter is organized as follows. Section II covers the limitations of current Internet architecture along with its research challenges and issues. Section III covers the basics of ICN, its main features, the prominent ICN architecture, the significance of ICN for loT, ICN-loT architecture, the suitability of ICN-loT architecture for edge computing and cloud computing. Research challenges in the implementation of the same have been also discussed. And Section IV ends the chapter with a brief conclusion.

Internet architecture and working

The present Internet architecture was designed in the 1960s and 70s when the users had the prime objective of file transfers and internet mail service. The suggested scheme has several technical critiques. Furthermore, the key issue was resource sharing that imposed serious challenges concerning communication among end systems. Thus the current internet architecture(TCP/IP) revolves around communication exchange between hosts and network devices. In TCP/IP Internet hosts can talk to each other by the communication pipes as it is established between them. It serves the client-server application model very efficiently.

With the advent of new technologies primarily at the network core and by the access networks, the bandwidth availability has increased and the user access costs have also reduced. This has resulted in flooding the internet with multimedia applications. And the popularity of multimedia applications with the cost dilution has thronged the Internet with millions of users. Moreover, every year millions of new users are adding up. As predicted by Cisco, over the next 5 years global IP traffic will increase nearly threefold. Remarkably, in 2016 IP video traffic was 73 percent of all consumer internet traffic and by 2021, it will reach 82 percent [17]. As a consequence, applications like video-sharing websites and file-sharing peer-to-peer (P2P) systems evolved. It indicates that the content distribution on the Internet has evolved from a textual information system to a multimedia information system. Thus, the users are more interested in content rather than the content source. Hence, modern applications are content-oriented. But the protocol stack is based on content location and TCP/IP provides no unique solution to this problem. Therefore, the development of network infrastructure that supports content distribution with high availability is the need of the hour. Hence, some partial solutions like Content distribution networks (CDNs) and P2P networks have been proposed for content distribution. And they led the foundation for some widely successful applications like BitTorrent and Akamai. However, both CDNs and P2P systems operate as overlays and the underlying network topology is not taken into account to improve the content distribution efficiency [108].

Research challenges and issues

Following are some of the major research challenges and issues in modern Internet architecture:

  • • On a global scale, the modern Internet is working as a packet-switched network. And Internet Protocol(IP) is designed to forward the packet using best-effort service. Thus the contents are distributed with no performance guarantee as there is no resource reservation or service differentiation while packet forwarding.
  • • It compels the users to know the content location as the source host needs to include the IP address of the destination host in the packet header for communication.
  • • Moreover, the current Internet uses the client-server model in which one point-to-point communication channel is established between one client and one server. Hence, if several users request a particular content hosted by a server, multiple point-to-point channels are needed to be established and one copy of the same content is sent over each of the channels. Thus, popular contents reduce the efficiency of content distribution in terms of bandwidth. Thus scalable forwarding mechanisms are required for large scale content distribution applications. [3]

example, Dynamic redirection in Domain Name Systems (DNS) and HyperText Transfer Protocol (HTTP) is used by CDNs but does not guarantee the persistence of content. Also, the content delivery time is increased as queries to centralized structures are needed to change content location.

• Content authentication and secure communication are also needed for content distribution applications. At present, a secure channel is provided between the source and the destination host instead of securing the content itself. And as a result, additional messages and process overhead are introduced. Moreover, content security depends on the trust of the host that stores the content and also the connection established between hosts.

Thus, mainly three characteristics of current Internet architecture are considered as hurdles to content distribution. There is no guarantee of (i) end-to-end security, (ii) quality of service, and (iii) no scalable forwarding mechanisms.

Hence, modifications are required in the current internet architecture which takes into consideration issues like content delivery, location efficiency, and content availability. And the main goal of Information-centric networks (ICNs) is to fulfill these requirements [108].

Information Centric Networks (ICN)

Information-centric networks (ICNs) introduced a paradigm shift in the Internet model. ICN lays stress on content rather than the location which allows the content to be delivered actively by the network. Content naming, name-based routing, in-network caching are some of the innovative concepts employed by ICNs.

Important terminologies used in ICN

This section covers the important terminologies used in ICN:

  • Information-Centric Networking (ICN): ICN is a concept for communicating in a network that provides accessing named data objects as a first-order service.
  • Named Data Object (NDO): In ICN, NDO is the terminology used for "the addressable data unit that represents a piece of information or a collection of bytes". And since a name is bound to each data object it is termed as Named Data Object. Different concepts are used in different ICN approaches for mapping Named Data Objects to individual units of transport eg. chunks and segments.
  • Requestor: In an ICN the entity that is sending the request for an NDO to the network.
  • Publisher: In ICN the terminology "publisher" is used for that entity that publishes an NDO to the network, and the corresponding request for that NDO sends to that publisher [206].

  • [1] 'The literal meaning of the term "information" is data converted into a more useful or intelligibleform. For, eg. host requests for the particular information, eg. date-of-birth of Xxx. And he gets inresponse a text content "19-12-1970".
  • [2] Here the term content refers to text, images, audio and video content.
  • [3] In addition to that, the present-day Internet architecture also experiencesdifficulty in content persistence, scalability, availability, and security. For
 
Source
< Prev   CONTENTS   Source   Next >