Definitions Editor’s note: this clause is currently empty 4 Abbreviations and acronyms

Figure 1 Draft architecture of 5G mobile network

An Information Centric Network is usually made up of content producers, content publishers, content replicas, and content consumers. A producer generates a piece of content, such as a document, photo, movie, or web page. It may have its own digital rights management (DRM) attached by the producer. A publisher packages a piece of content for use in the network. This may include pre-encoding the content to certain formats and names and signing them with a network identity. A replica distributes content from a publisher. A consumer fetches content via network names from replicas. The download process at a consumer understands the inherent security offered by the ICN, which usually allows authenticating every packet via direct signature or implicit hash chain from the publisher. This is different than today’s security model, where authenticity derives from a secure connection to a replica. In the simplest configuration, one entity is a producer, a publisher, and a replica for its content.

Figure . Typical ICN architecture

The 5G network architecture should yield a framework that allows the reshaping of the economic foundation upon which the mobile network operates such that the new network can both efficiently serve the emerging use cases and markets (eg IoT) and support current usage models (eg those that are video related) at substantially lower cost points. Our expectation is that properly crafted, the impact of a solution that successfully addresses these challenges would extend well beyond the bounds of the mobile network. In the following paragraphs, we identify characteristics of the new solution space that address the perceived deficiencies in current network design:

1. 
   Mobility – current networks base packet forwarding decisions on location-based labels that identify the point-of-attachment of the destination host on the network. Whenever the host point-of-attachment changes (ie a mobility event), a means for updating the association of the host to its network-addressable label is required, one that also preserves the socket-based connections that mobile applications are bound to. Mobile networks employ tunnel-based solutions for this purpose. Tunneling ensures reachability of mobile hosts via a layer of indirection that allows the mobile host point-of-attachment label to change while preserving the socket structure. One can view the necessity of anchoring gateways required to implement and dynamically manage tunnel-connections as a direct consequence of network design that builds on a foundational layer that was not architected to support mobility. The current design burdens mobile traffic with an incremental cost that is traceable to this architectural consequence, one that requires mobility be treated as an explicit feature rather than as an emergent property of the basic design. The 5GIMT-2020 network should remedy this deficiency and treat mobility as a first class citizen in its design and eliminate the need for tunnelling and the associated complexity and cost that accompanies it.
   
2. 
   Privacy and Security – Privacy has always been a major concern for mobile networks. The RAN has employed link-encryption by default since the second generation standards and LTE commonly employs IP SEC in backhaul network between the eNB and PDN gateway for protection along those connections. The importance of guaranteeing privacy on all Internet services goes beyond the specific radio mobile data network or the all IP core; it includes the communication service to all of its components: over-the-air transmission, end-to-end IP, HTTP and web service that might employ user trackers or store private content in clear text at the server end. The current solution guarantees authentication, integrity and confidentiality by delegation to the transport layer on an end-to-end basis using TLS which may serve a short term objective but at a cost of significantly impairing long-term flexibility, reliability and manageability. The use of TLS implies a significant cost that is quantifiable²⁷ although it is the default solution in the HTTP/2.0 standard despite the criticisms about its weaknesses as a technical solution to the privacy problem.
   The issues regarding user privacy and data authenticity, integrity and confidentiality are crucial to get defined accurately and addressed appropriately in the IMT-2020 basic design. It is important to recognize from the outset that different applications, usage models, market models, and so on will have diverse risk exposures that beget different security requirements. It is unlikely we can successfully anticipate all potential combinations and it is similarly unlikely that solutions that trend towards worst case coverage will be economically acceptable. This suggests an architectural design that enables flexible application of security elements appropriate to the application and usage requirement. The decomposition of security in component functions (eg data authenticity, integrity, confidentiality) such that each can be individually addressed is a desireable capability that allows the flexible creation of purpose-directed security and privacy solutions that can address differing risk profiles successfully within a cost-effective ecomomic model.
   
3. 
   Transport Efficiency – Reliable and stable transport over current networks is managed by a distinct protocol layer (eg TCP, SCTP, etc) that operates between communicating end-points. The end-to-end action of these protocols treat the composition of links that form the communication path in aggregate that lead to the use of abstractions for congestion detection and remedial action that lead to poor behaviour when applied to composite paths that include wireless links. The anomalously erratic behavior exhibited by TCP in mobile networks is traceable to the (mis)interpretation of variations in link bandwidth and latency as a sign of network congestion resulting in inappropriate backoff responses, unnecessary data retransmissions, increased download times and more generally inefficient usage of wireless access and network resources. An architectural approach that enables hop-by-hop congestion detection and management for accurate interpretation of link behaviour and initiation of appropriate responses is desireable. This is particularly important in the case of wireless links that employ (multiple access) link layer protocol may are highly adaptive and designed to deliver high spectral efficiency. The objective is to have intrinsic congestion and flow management mechanisms that account for the unique characteristics of the wireless link.
   Current tunnel-based mobility transport also frustrates the use of in-network storage and caching of content within the mobility network. Caching should be viewed not simply as a means for improving delivery efficiency of popular content but also in its roll in facilitating retransmission and reducing latency when reliable delivery is required. The 5GIMT-2020 mobile network design should allow the flexible use of in-network storage as an architectural component.
   
4. 
   Latency Sensitivity: 5GIMT-2020 places stringent requirement such as 1-10ms for certain class of applications; also in general any applications benefits from reduced response time through better throughput and faster service logic execution. Though delays such as 1-10ms may be very difficult to achieve in software driven implementations, in normal situations distance of the consumer from the service contributes significantly to the end-to-end application delay. Though there are also tradeoffs to manage the distributed service instances, benefits are also realized as overall load of the service is amortized among the distributed instances. ICN features such as naming, name-based service discovery, name-based routing allows application to discover the closest service points with which it can transact. In certain situations, like IoT, these services can be placed at extreme edges of the network such as over public infrastructure to address the need of mission critical applications.
   
5. 
   
   
6. 
   Bandwidth Efficiency: The estimated increase in wireless capacity is expected to be 1000X for 5GIMT-2020. ICN can help this situation in multiple ways: 1) eliminating redundant data transmission considering multicasting and caching is integral feature of ICN; 2) cheaper computing and storage will enable significant intelligence end point and infrastructure elements well suited for ICN to exploit, here ICN enabled application and service interaction can be localized operating over both wired, unlicensed and licensed spectrum, thus offloading the backhaul from any control or data plane overhead due to these interactions ; 3) also tighter integration of ICN with MAC layer to enable multicast and broadcast of data objects can improve bandwidth efficiency of the wireless resources, particularly helpful for very popular contents and flash crowd situations.
   

- object security

- latency minimization

- hop-by-hop dynamic congestion control

- multipath transport

- optimized transport reliability

- mobility

- …