Abstract—Enabling content-centric network (CCN) features over a cellular network can significantly reduce mobile data traffic and support upcoming 5G. However, traditional cellular networks follow a host-centric network architecture rather than the content-centric architecture. Moreover, supporting CCNspecific features, such as multicast and broadcast of a content to multiple users, searching a possible content source in a wireless environment is extremely challenging due to constraints of limited power and user mobility. The primary question remains, how to distinguish between CCN packets and normal cellular network packets? Is it possible to enable this features at router level? In this article we explore several technologies and packet formats to support coexistence of CCWN and traditional cellular communication.
I. INTRODUCTION
Annual mobile traffic is growing exponentially and suggestively the growth rate was 63\% [1] in 2016. Rapid proliferation of tablets and mobile devices suggest even greater future growth prospects but finiteness of the radio spectrum forwireless communication makes it highly challenging to regulate exabytes of data traffic successfully. Variousinnovative technologies are being developed to support the traffic, and cache enabled content centric networking is one of the most promising technologies [2]. Video delivery companies (e.g., YouTube, Netflix) already use simple forms of popularity based in-network caching in today’s Internet to improve user performance. These video delivery applications primarily determine the popularity of multimedia content based on parameters such as release date, viewership of past series of a show and push popular content to the network edge [3]. Recent performance analysis works for content centric networking (CCN) suggest that caching popular content within the network significantly improves network’s performance.
Typically, CCN is a multicast enabled architecture, where multiple users requesting the same content are served together from a nearby cache instead of the original content server. Multicast manner for content distribution is a simple method to alleviate traffic load in both network and content sender regardless the number of receivers. Moreover, multicast routing features the following relevant aspects.
- Data is pushed to receivers, supporting multiple transport streams in parallel and eliminating the need to ask for content changes.
- Data distribution supports immediate in-network forwarding and is suitable for efficient, scalable real-time streaming in particular. This mainly covers the use case of real-time date dissemination without storage or caching requirements.
- The multicast model contributes many-to-many communication, which is valuable whenever information is created at distributed origins. Multisource communication using a single name faces strong conceptual difficulties in unicast-based CCNs.
- Multicast group communication enables rendezvous processes, as publishers and subscribers remain decoupled and unknown to each other.
Although multicast feature of CCN is extensively explored in literature [4], [5], combination of CCN and multicast scheme in a cellular network is remains a bottleneck because wireless multicast scheme at MAC does not guarantee complete packet delivery [4]. Moreover, distinguishing IP traffic and CCN traffic, i.e., differentiating a CCN packet and a traditional IP packet at network level is a challenging problem for future coexistence of CCN and wireless cellular network. Therefore, in this article we explore how to standardize multicast routing and multiplex CCN and IP network features in a cellular network.
A. Paper Organization
The rest of the paper is organized as follows. We provide an overview of existing coexistence problems in Section \ref{coexist}. Network architecture and different features of CCWN is discussed in Section \ref{net_arch}. Thereafter, we propose some of the possible technologies to support the coexistence. Finally we conclude the article with open research problems and future directions
II. COEXISTENCE PROBLEMS ARE NOT NEW
Coexistence challenges in communications and networking are present from the very beginning of standard transitions. At outset, the inevitable transition from IPv4 to IPv6 to avoid the IP address exhaustion and routing scalability employs the IPv4/IPv6 coexistence [6], [7]. Initially IPv6 was designed with no backward compatibility with IPv4, but the vast network resources and services were still using the IPv4. Hence, IPv4 and IPv6 will coexist for a longer period. To achieve this heterogeneous traversal, principles of tunneling was introduced [8]. To deliver IPv4 packets over IPv6 network in the middle, first, we need to encapsulate IPv4 packet into the new IPv6 payload then again at the next tunneling endpoint, we need to decapsulate it to get back the IPv4 packet.
Similarly, coexistence problem is also observed between Long Term Evolution (LTE) from 3GPP and WiFi. It came to picture when LTE had to expand to the unlicensed bands between 2.4-5 GHz to support export exponential growth of mobile users, and incidentally these were the unlicensed bands where WiFi channels were allocated previously [9]. The fair and friendly LTE-WiFi coexistence can be achieved by deploying new protocols e.g. LTE Unlicensed (LTE-U) and Licensed-Assisted Access (LTE-LAA). The first one uses the duty cycle-based approach and carrier aggregation where the latter one uses listen before talk (LBT) [10], [11].
In this proceeding, the coexistence of TCP-IP with the Information-Centric Networking protocol (CCN) also imposes a major issue for carrying TCP traffic over CCN transparently. In order achieve a fair coexistence, a protocol translation along with tunneling have been proposed [8]. Tunneling will aid to encapsulate TCP segments to interest packets or decapsulate from data packets. On the other hand, protocol translation helps to coexist the functionalities of pull-based CCN and push-based TCP model.
Basic methodologies used to attain a fair coexistence between a newly adapted technology and already available cellular technologies are an introduction of inter-proxy protocols in order to make two technologies synchronized without having any interference or collision, and most importantly disrupting and users experience.
III. NETWORK ARCHITECTURE
The architecture of a typical wireless cache-enabled network is illustrated in Fig. 1. At the bottom of the hierarchy, there are access points with smaller coverage area, called eNodeB. From CCN point of view eNodeBs can be called edge caches, which are connected with the users over a wireless medium. SBSs are connected to to the servers via a multilayered backbone network. The backbone network consists of wired routers, termed core caches. Now depending on the user activity, cellular data traffic from Macro BS or enodeBs can be IP based or CCN traffic. If the content is available at another user or eNodeB, then request is generally routed to that node, thus leading to D2D and M2M communication. Therefore, M2M, D2D communications and communication of WiFiAP in the unlicensed band coexist along with this. Now IP requests and CCN requests that are not served by a edge cache is aggregated at the aggregator.
A question immediately comes to mind, how to packetize CCN requests and CCN related traffic in traditional IPv4 and IPv6 to aggregate all of them together using aggregator? In next section we will discuss possible solutions for this challenge. FOr the time being let us assume that CCN requests can be encapsulated in a IP packet. Now all the aggregated IP data-traffic not found in edge routers or nearby server, is searched in the interconnection of routers in the IP network following the shortest path algorithm. Here efficient caching and routing mechanism are extremely important to properly utilize the network caches. A comprehensive study of caching strategies is discussed in [12], and a study of routing and content naming strategies is discussed in [13].
Fig. 1. Typical Network Architecture of a CCWN
If requested content is not found at the IP network, requests enter into backbone network, first in the network cloud, which typically consists of backbone mainframes or core caches in CCN terminology. Finally, if the requested content is still unavailable then the requested is forwarded to the data center (global content server) through the gateway. Basically, a CCN request traverses the distance between UE and server in search of alternative content provider, whereas traditional IP requests are directly moved towards the server.
IV. COEXISTENCE TECHNOLOGIES
In this section we discuss several possible technologies to bridge the gap between CCWN communication and traditional cellular communication. Primary problem of their coexistence is distinguishing between IP packets for traditional cellular communication and CCWN packets for cache networks, and thereafter, routing these different class of packets accordingly. So, the problem can be reduced to separation of routing plane and data plane. Separation of routing and forwarding planes also allows the incremental deployment of new data planes over the programmable compute, storage, transport infrastructure [14]. CCN can work in tandem with the software-defined network (SDN) and Network Function Virtualization (NFV) in order to achieve this multicast-driven networking framework. The deficit of inter-domain IP multicast can be resolved by this content-centric publisher/subscriber model [15]. These salient features of CCN can be integrated with IP intra-domain multicast to achieve forwarding efficiency using the shortest path routing. Existing receiver-driven TCP protocols can also be applied for CCN while generating interest packets and sending the receiver buffer capacity [16]. Relying on the receiver buffer size, CCN packets can be segmented into chunks and each chunk can be named differently [17]. Considering the humanreadable hierarchical names, the different chunk naming for the same content packet can be realized just by changing the leaf nodes’ names which makes the parsing of content names for various chunks easier. After different freshness period of a particular content packet, if users ask for updated information, the custodian can send the particular chunk file which is modified rather than sending the entire packet and wasting the network bandwidth. So, a possible solution to support the routing of CCWN and IP packets is employing SDN as an overlay.
While traversing through the network, CCN packets need to be discriminated from the existing IP packets. In the protocol field of IP header, we can have one protocol option for CCN packets. Since CCN packets are also divided into interest and data packets. We can include one option for each in the Type of Service (ToS) field. This ToS field rarely used in TCP header. If the protocol file corresponds to CCN and ToS field corresponds to an interest, the edge routers can get the content name from the destination address field of the IP header and then they will perform hash mapping to get the custodian for the content. Queuing of multiple interest packets for the same content can also be resolved by the ToS field and destination field which contains the name of the content packet. The Nonce field in CCN packets to resolve the looping of the same packet can be unraveled using the TTL field in the IP header. Interest Lifetime and forwarding hints in CCN packets can be embedded inside the options field of IP header. These fields can affect the interest packet routing towards the custodian [18].
The exploitation of already available TCP-IP header format for CCN supports to attain a fair coexistence between TCP-IP and CCN traffic with modifications in the interpretation of few header fields. These modifications in interpretations can be easily resolved by programmable compute and decision making (control) plane developed with the aid of SDN over the top of CCN and traditional cellular traffic.
V. CONCLUSION
From the beginning of communication and networking, it has been a pursuit of bridging the gap between multiple technologies. Coexistence problems come into scenario whenever this heterogeneous mixture of technologies are required. In this article we discussed possible challenges for coexistence of CCWN and traditional cellular communication, due to the difference in their routing mechanisms and several other salient features of CCN. We studied how SDN can be a useful technology to overcome the routing challenge. We also discussed how switching the packet formats can solve the naming challenge and discriminating between IP packet and CCN packet. However, still there are several open research problems, such as, implementation of SDN over a cellular network, especially controlling the edge connections using SDN require a lot of infrastructure. Moreover, how additional processing delays impact the overall user experience? These are the questions that require further exploration.