I have given a keynote speech at the IEEE MASSAP Workshop at ICC 2015 in London. The talk is on wireless massive and ultra-reliable communications, which are seen as two new modes that will be featured in 5G. There is, of course, a technical part in the talk, but there is also a part which argues why the research on wireless and communication theory is still vital. The slides can be found here:
Wireless 5G systems will not only be “4G, but faster”. One of the novel features discussed in relation to 5G is Ultra-Reliable Communication (URC), an operation mode not present in today’s wireless systems. URC refers to provision of certain level of communication service almost 100 % of the time. Example URC applications include reliable cloud connectivity, critical connections for industrial automation and reliable wireless coordination among vehicles. This paper puts forward a systematic view on URC in 5G wireless systems. It starts by analyzing the fundamental mechanisms that constitute a wireless connection and concludes that one of the key steps towards enabling URC is revision of the methods for encoding control information (metadata) and data. It introduces the key concept of Reliable Service Composition, where a service is designed to adapt its requirements to the level of reliability that can be attained. The problem of URC is analyzed across two different dimensions. The first dimension is the type of URC problem that is defined based on the time frame used to measure the reliability of the packet transmission. Two types of URC problems are identified: long-term URC (URC-L) and short-term URC (URC-S). The second dimension is represented by the type of reliability impairment that can affect the communication reliability in a given scenario. The main objective of this initial work on URC is to create the context for defining and solving the new engineering problems posed by URC in 5G.
The full article is available here.
The IEEE 1st Int. Workshop on D2D and Public Safety Communications (WDPC) took place on April 6, 2014 in Istanbul, in conduction with the IEEE WCNC conference. Petar was one of the keynote speakers, presenting the talk “What is in for D2D in 5G wireless and how to support underlay low-rate M2M links”. The slideshow of the talk can be downloaded here. The full program can be seen at:
The METIS project has released a document describing initial system concepts for the 5G wireless system. Our research group is actively involved in multiple aspects of METIS and in this document I have participated in the capacity as a driver of the topic Ultra-Reliable Communication (URC). In addition to the common approach to work on detailed wireless technologies, METIS has also defined multiple overarching topics, termed Horizontal Topics, URC being one of them. A horizontal topic covers the system-level aspect of a certain new feature that is envisioned for 5G wireless systems and this document is an initial report on the system concept per horizontal topic. The document is available here and the executive summary is given below.
The overall purpose of METIS is to develop a system concept that meets the requirements of the beyond-2020 connected information society and extend today’s wireless communication systems to support new usage scenarios. This is an immense task, residing at a level that is different from the level at which detailed technical innovations are created. Therefore, the task of overall METIS system concept has been segmented into system concepts related to five Horizontal Topics (HTs): (1) D2D – Direct Device-to-Device Communication, (2) MMC – Massive Machine Communication, (3) MN – Moving Networks, (4) UDN – Ultra-Dense Networks, and (5) URC – Ultra-Reliable Communication.
This document provides a first view on the system concepts associated with each HT and indicates the steps and directions towards integrating them into an overall METIS system concept in the remaining time frame of the project. Each HT creates a context for applying and optimizing the Technology Components (TeC) in the Work Packages (WPs). For example, Massive MIMO is a generic technology that can give rise to innovations applicable in UDN and supporting the performance requirements specified by URC. The five HTs are not independent and their interaction is the basis for creating the architecture that is capable to achieve the main objectives for METIS [MET13-D11]:
- 1000 times higher mobile data volume per area,
- 10 to 100 times higher typical user data rate,
- 10 to 100 times higher number of connected devices,
- 10 times longer battery life for low power devices,
- 5 times reduced E2E latency.
These objectives have to be met at a similar cost and energy consumption as today’s networks. The research challenge is amplified by considering that the wireless scenarios in 2020 will feature communication modes and services that are not merely “more and faster of what we have today”. As an example, the vehicles of the future that are interconnected with very high reliability and low latency, improving the efficiency and safety on the road. The METIS system will respond to the requirements for improved: 1) efficiency in terms of energy, cost and resource utilisation than today’s system 2) versatility to support a significant diversity of requirements, e.g. connections in Gbps from few devices vs. connections in kbps from many machine-type devices, inclusion of moving networks vs. statically deployed sensors, etc. 3) scalability in terms of number of connected devices, densely deployed access points, spectrum, energy and cost.
The concepts presented in this document are clearly demonstrating the capability of the HTs to channelize the technical innovations towards creating the overall METIS system. The highlights for the system concepts of the individual HTs are given as follows:
- The HT D2D concept addresses the utility of the local exchange of information among the devices and creates a framework for solving the associated technological challenges. Putting D2D connectivity as a basic architectural element in the 5G system, rather than having it as an add on to an already existing architecture, leads to multiple benefits: increased coverage (availability and reliability), offload backhaul (cost efficiency), provide a fall-back solution (reliability), improve spectrum usage (spectrum efficiency), typical user data rate and capacity per area (capacity density), and enable highly reliable, low-latency Vehicle-to-Infrastructure (V2X) connections. Efficient D2D operation critically depends on interference management, resource allocation, efficient relaying for coverage extension, etc.
- The HT MMC concept contains the technologies for radio access that will be capable to support an unprecedented number of devices. They are segmented into three types of radio access: (1) direct access, which devices transmit directly to the access node; (2) access through accumulation/aggregation point; and (3) machine-type communication between devices. The TeCs used to support these types of access fall in several categories: overlay of multiple transmissions (by means of quasi-orthogonal random access, sparse coding, successive interference cancellation, and reuse of resource by short-range links), smart pre-allocation of resources (persistent scheduling), techniques to lower the sync requirements and context/service-aware configuration of the radio access.
- The HT MN concept introduces innovative directions for the future relationship between vehicles and wireless communications. Three clusters are defined: (1) MN for mobility-robust high-data rate communication links (MN-M), to enable broadband as well as real-time services in mobile terminals and moving relays; (2) MN for nomadic network nodes (MN-N), to enable a flexible and demand-driven network deployment; (3) MN for V2X communications (MN-V), to enable reliable and low-latency services such as road safety and traffic efficiency. While the MN-M cluster represents an evolutionary improvement of the existent technology addressing highly mobile scenarios, the MN-N and MN-V clusters introduce a paradigm shift in the usage of mobile communications.
- The HT UDN has defined a core specific concept optimized for the potential stand- alone operation of a layer of ultra-densely deployed small cells. Beside considerations on a new spectrum flexible air interface, it foresees a potentially tight collaboration of nodes w.r.t. resource allocation coordination, a fast (de-)activation of cells and inbuilt self-backhauling support. An extended UDN concept offers additional performance improvement by: 1) Context awareness for mobility, resource and network management, 2) inter-RAT/ inter-operator collaboration, 3) tight interaction of a UDN layer with a macro layer holding superior role in control and management functions over common area, and 4) macro-layer based wireless backhaul for flexible and low- cost UDN deployments.
- The HT URC system concept targets operation modes that are not present in today’s systems. URC-L (Long-term URC) targets the following operation mode: when not possible to operate at the peak rate, provide reliable moderate rates to all users instead of failing some of them. URC-L will be instrumental for the cloud-based services of the future. URC-S (Short-term URC) aims to guarantee latency despite the competition from multiple users and varying channels and will be critical for e.g. V2X connectivity. URC-E (URC for Emergency) aims to provide minimal guaranteed connectivity upon infrastructure damage. A generic URC toolbox consists of: spectrum allocation and management, robust PHY mechanisms, signalling structure and interface management, Multi-RAT and reliable service composition.
Based on the current view on the system concept, METIS has selected two technology components for implementation test-beds. The technology components are “Direct network controlled device to device communication with interference cancellation” to be implemented on the Radio Resource Management test-bed, and “FBMC/OQAM new waveform” for the Digital Base-Band test-bed.
The HT-specific concepts will be integrated towards the overall METIS system, which will contain new air interfaces as well as the evolved versions of today’s systems. In order to deal with such a level of complexity and support the required reliability/scalability, the METIS system will feature Software-Defined Networking (SDN), Network Function Virtualization (NFV), and Self-Organizing Network (SON) technologies.
The next steps of the work on the METIS system concept in the project include: further integration of the specific HT concepts with the METIS architecture, further positioning of METIS technical goals by the evaluation criteria defined for 5G systems and establishment of a technology roadmap for the deployment of the METIS 5G system.
We were part of a team that recently wrote an article “Five Disruptive Technology Directions for 5G”. The article was featured in MIT Technology Review and highlights the following figure. Although the figure is conceptual and with no ambition to provide precise numbers, it is interesting to see the reasoning behind the regions. The x-axis is represents the number of users in a cell and the y-axis represents the data rate per device.
- R1 – Describes the operating region of today’s wireless wide-area cellular systems. As the number of users increases in the system, the data rate per user decreases.
- R2 – It is the region targeted by the wireless research efforts that aim to increase the data rates, such as mmWave, network densification, massive MIMO, etc. Note that the upper part of this region decreases as the number of users increases, but slower compared to R1. This aims to highlight that many of the new broadband high-spectral efficiency techniques are inherently designed to serve efficiently multiple users, such as: interference alignment, wireless network coding, improved spatial reuse in mmWave wireless, etc.
- R3 – It is the region of massive M2M communication with relatively low data rate and therefore termed lowband communication. The required data rate is able to continuously decrease with the the increase of the number of connected devices, but not go abruptly to zero as the system becomes more congested. It is assumed that each device transmits sporadically and the y-axis represents the average data rate of the device. The declining becomes progressively faster due to the fact that short packets, sent by many devices, are affected by the signaling overhead. To see why this region represents a research challenge, we can think of this as follows: it is relatively easy to send 10 Mbps from each of the 10 devices, but it is much more difficult to send 10 kbps from each of the 10000 devices. The problem is that in the latter case the overhead of the protocols, collision waste, etc. start to dominate the performance.
- R4 – It is the region of lowband communication, but very low latency or ultra-high reliability. These features are not plotted on the figure. This region represents the connectivity required in e. g. car-to-car communication (low latency, high reliability), public safety, critical control of small-scale M2M installations, etc. To see why this region represents a research challenge, we can think of this as follows: it is relatively easy to send support 10 Mbps to a device during 95% of the time, but it is very difficult to guarantee 10 kbps during e. g . 99.999% of the time. If the 5G wireless connections are able to deliver such a reliability of a stringent latency requirement, then the number of new services will proliferate, as people and industries will start to trust the wireless connection as being a “true” cable replacement.
- R5 – region that cannot be reached due to fundamental limits of physics, information theory and networking theory.
We believe that a concise description of what 5G will be is given by the following:
Instead of being only a broadband “4G, but much faster” system, 5G will also include lowband communications, represented by the regions R3 and R4.