Internet Engineering Task Force LiOu Internet-Draft Sun Wujian Intended status: Standards Track NDSC Expires: June 10, 2011 December 10, 2010 Proactive Channel Handoff in wireless cognitive networks draft-liou-sunwujian-manet-proactive-handoff-00 Abstract Driven by government policy and advanced radio technologies , opportunistic spectrum usage plays an ever-increasing role in the next generation communication system. Due to the random characteristics of primary users, it is impossible to fully describe the activity feature in terms of time and space, rendering opportunity usage unsafe. In order to minimize interference to primary users, a history knowledge based algorithm is proposed in this work, which compound historical statistics with current information. Random early detection (RED) algorithm, which proves to be efficient in active queue management (AQM) for switching fabrics, is introduced to help select the desirable channel. In order to further suppress the collision probability, all the channels are sorted according to availability. Simulation results show that, compared with completely random spectrum selection method, the proposed RED-based channel access method gives better performance in terms of channel selection delay. Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." This Internet-Draft will expire on June 10, 2011. LiOu & SunWujian Expires June 10, 2011 [Page 1] Internet-Draft Proactive December 2010 Copyright Notice Copyright (c) 2010 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 3 2. Proactive channel handoff method . . . . . . . . . . . . . . . 3 2.1. Proactive channel handoff introduction . . . . . . . . . . . 4 2.2 Random Early Detection . . . . . . . . . . . . . . . . . . . 4 3. System Model . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.1. Statement of the Problem. . . . . . . . . . . . . . . . . . 5 3.2. Channel reliablity and average remaining time. . . . . . . 6 3.3. Prediciton-based proactive channel selection. . . . . . . . 7 4. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 9 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 9 6. Security Considerations . . . . . . . . . . . . . . . . . . . 10 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 10 7.1. Normative References . . . . . . . . . . . . . . . . . . . 10 7.2. Informative References . . . . . . . . . . . . . . . . . . 10 8. Authors' Addresses . . . . . . . . . . . . . . . . . . . .. . . 10 LiOu & SunWujian Expires June 10, 2011 [Page 2] Internet-Draft Proactive December 2010 1. Introduction The rapid developments in wireless communication technologies and excessive growth of short-distance devices have resulted in ever- crowding spectrum usage. For the sake of static allocation of radio frequency, the entire spectrum band has already been handed out. Due to shortage of available licensed spectrum, wireless communication systems seem to encounter a bottleneck. However, according to authoritative reports of spectrum usage observation, at any given time and location, only a small proportion (less than 15%) of valuable spectrum is working, leaving much of the precious resources critically underutilized. Cognitive radio (CR) is proposed as a promising technology to cope with current spectrum scarcity problem by means of enabling cognitive users (or, secondary user) to dynamically adjust its operating parameters and thus dynamically reuse the spectrum which is always statistically underutilized by authorized users (or, primary users (PU)) in a intelligent and cautious manner. One challenging issue that CR networks inevitably encounter is channel selection when initiating a data link between the cognitive users. Compared to other hot topics of CR (i.e., spectrum sensing, spectrum management, and resources allocation), spectrum handoff problem receives little attention and is less investigated in the research area. 1.1. Requirements Language The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119]. 2. Proactive channel handoff method Spectrum handoff occurs when detecting the usual return of primary user. This leads to 1) interruption of the communication between cognitive users, 2) interference to primary users according to system mechanisms. Nearly all the current literatures simply assume that all the channels are well synchronized, whether by primary users (usually infrastructures of primary user network, such as GSM base station, etc) or just omitted such complicated scenario. Current channel scheduling methods concentrate mainly on sense- and-react approach, which means to randomly pick up a channel to access from preferable channel lists (PCL) if it were sensed idle at some given time t. This always leads to disruptions to both primary users and cognitive users. Proactive method in a sense-and -avoid manner instead seems more reasonable. However, due to the randomness of the reappearance of primary users, it is extremely challenging to perform fast, smooth and seamless channel handoff LiOu & SunWujian Expires June 10, 2011 [Page 3] Internet-Draft Proactive December 2010 for cognitive communicating parties, leading to inevitable performance degradation during a spectrum handoff. This problem becomes even more difficult in ad hoc networks where there is no centralized infrastructure to coordinate the spectrum mobility. 2.1. Proactive channel handoff introduction Currently, most channel selection algorithms take a reactive sense-and-react approach to perform spectrum handoff based entirely on the latest observations. One common way employed in relevant literatures is that cognitive users (CU) perform spectrum handoff as soon as perceiving the return of primary user .Borrowing the idea of proactive routing protocols designed for wireless Ad-hoc networks, proactive channel selection adopts a similarly active way that prepares one or more channel(s) before there is a need for communication parties. Although the merits of this approach are obvious, preparation of channels, or rather prediction of channel availability status is relatively difficult. Until very recently, only a few literatures focus on the spectrum handoff issue, among which fewer are concerned with proactive channel selection. But the idea is widely accepted of performing spectrum handoff and RF reconfiguration before a PR user reoccupies the spectrum band based on historical channel usage statistics, usually referred to as proactive channel selective approach. Also a comprehensive introduction is presented and a framework of proactive channel selection is briefly given without detailed analysis. A predictive model is proposed for dynamic spectrum access based on the past channel usage history. In , spectrum occupancy features and channel prediction model is proposed by means of the analysis of binary time series. In these few proposals, the multi-user network coordination problem is either not taken into consideration or purely assumed there exists a stable global common control channel (CCC), which in practice seems impossible due to the high dynamical nature of primary user. Nearly all the existing literatures pay little attention to the channel selection method employed in the cognitive users, nor does the residual time on each selected channel. 2.2. Random Early Detection Random early detection algorithm is first proposed in active queue management for gateways to keep the average queue size reasonably low while allowing occasional bursts of packets in the queue. Using a feedback mechanism, RED are capable of timely reconfiguration before congestion happens. According to, the average length of queue is: Qave=Alpha*Qave+(1-Alpha)*Qcurrent LiOu & SunWujian Expires June 10, 2011 [Page 4] Internet-Draft Proactive December 2010 In other words, the expectation value of the length of queue should be calculated according to the long history and latest observation. This means some unusual busty arrival must be dropped. Such a tradeoff between admission rate and buffer size have to be carefully chosen. Numerical results are shown in Figure 1. It can be seen that with Alpha big enough (more than 0.975), the average value is less than 50 even if instantaneous arrivals reaches 200. This fits much well with the idea that impulsive congestion is smoothly averaged at the cost of dropping some packets. The parameter Beta=1-Alpha is called time constant in term of low-pass filter. This is similar to the time constant in low-pass filter, which is an exponential weighted moving average (EWMA): This is the same with the conclusion of. The determination of is the key of using RED. To the best of our knowledge, this is the first paper that incorporates the random early detection algorithm in the spectrum handoff design. 3. System Model The cognitive networks contain two kinds of nodes, or rather users, namely primary users and cognitive users. The resource of cognitive users comes from the primary users, in a manner of opportunistic usage. In this work, resources are actually frequency channels authorized to licensed users. Each channel is of two states, busy (occupied) or idle. So, an alternative renewal process can be used to model the licensed channel. The model of a primary networks is shown Figure 2. In this work, we assume that each cognitive user is equipped with two radio units, one for channel statistics (both long and near history) and the other for data transmission (Sending, receiving and sensing). The one for channel statistic is further assumed to be wide enough, capable of scanning all the licensed channels among the cognitive networks. 3.1. Statement of the Problem As assumed above, each channel is modeled as an alternative renewal process. Then the aim of cognitive users is to opportunistically utilize the idle fraction (mostly large part) of time. Assume that the beginning state of channel is busy (idle is also acceptable and it does not matter in the long run), it lasts a random time, say X1, then it changes to the contrary state, lasting a random time, say Y1, then goes to busy state again with random duration X2, and so on and so forth. In this sense, random vector{Xn,Yn} is obtained.Xn,Yn is independent of each other.Without losing LiOu & SunWujian Expires June 10, 2011 [Page 5] Internet-Draft Proactive December 2010 generality, we identify each channel a different parameter, i.e., the random variable sequence {Xn,Yn} are subjected to different parameters, noted as (Xni,Yni),1<=i<=N,n>=1.The distribution function of Xi and Yi is expressed as FXi,FYi. The available channel for cognitive users is the licensed channel with idle state. This work aims to pick up a more reliable channel in terms of remaining time of idle phase. At any specified time, there are several (even many as stated above that 85% of total channels are unused) channel for cognitive users to select. Our questions turn out to be that choose a channel with more idle time left. 3.2. Channel reliablity and average remaining time Based on the renewal theory, given the existence of E(Y*Y), namely E(Y*Y)