A Survey on Channel Scheduling in WLAN

A Survey on Channel Scheduling in WLAN

Chaithra V [1] [1] PG Student, Dept. of ECE, GSSSIETW, Mysuru

M V Sathyanarayana [2] [2] Professor and Head, Department of ECE, GSSSIETW, Mysuru

causes retransmissions and even fra mes drops. While OCF may mitigate the impact of these interferenc es, not all collisions can be a voided. Moreover, frame collisio ns may make the APs decrease their PHY tra nsmission rate , whic h results in performa nce de gra da tion. On the other band, if stations experie nce EN interfere nce , the ir assoc iated APs can benefit from simultane ous transmissions and achieve improve d throu gh put. However, if an AP senses the signa l of other APs, it defers its transmission and fails to e xploit the EN interferenc e .

To reduce interfere nce , proper channe l sc heduling is needed In this surve y pa per, we surve y the solutions for cha nne l scheduling in WLAN and ide ntify th e proble ms in these approa c hes.

1. SURVEY

   A Perro in his work[ I] Present eda unique measure ment study of WiFi experience in home environments through the le ns of WiSe APs across

   30 homes (§2.3). Through our measure ments we observed that while most of the WiFi c lie nts experience moderate to good performa nc e, poor performanc e plagues these e nvironments about

   2.1% of the time . The major cause of poor network

   performa nce (a irtime , signal stre ngths) was depe ndent on the e nvironm e nt. Some APs experie nced short periods of high impa ct

   patterns, and bursty traffic mixes. We also show that, in a multi-cell network, however, the collision rate increase s signific antly. And a lso Providing novel insights into the behavior of TCP in multi cell WLANs. Due to TCP flow c ontrol the number of bac klogged sta tions equa ls twice the number of access points, meaning that network effic ie ncy is determined by the nwnber of interfering acc ess points, not the nwnber of clie nts.

   R Gummadi in his work [4] ide ntify se veral factors that le ad to these vulnera bilities, ranging from MAC la yer driver imple me ntation strategies to PHY layer radio fre que nc y implementation strategies. Our results further show that these factors are not overcome by simply c hanging 802.11 operational para meters (such a s CCA threshold, rate and pac ket size) with the exception of freque ncy shifts. This le a ds us to e xplore rapid channel hopping as a strategy to w ithstand RF interfere nce . We prototype a channe l hopping design using PRISM NICs, a nd find that it can sustain throughp ut at le ve ls of RF interferenc e we ll above that needed to disrupt unmodified lin ks, and at a reasonable c ost in terms of switc hing overheads.

   Y. Lee in his work [5] proposed \VL AN servic e design process in following steps: Estimation of the demand area map, Se lection of candidate locations for APs, Signal me a sure me nt a l the de ma n d point in

   1

   interferenc e (81 % degradation] from externa l the service area Decide APs without channe l

   sourc es (e .g., microwaves). Also, ma jority of APs at homes tend to have static c hanne l configurations over time , indic ating that these APs do not adapt to interfere nce or c ontention e xperie nced by APs due to external sourc es.

   M A Ergin in his work[2] proposed the ana lysis of system performance using a rea listic TCP dominate d traffic mix in multi-ce ll WL ANs. Results show that a single-ce ll network re ma ins remarka bly robust e ven with 125+ c lients: the collision rate re mains low. This extends Choi et.a l. 's e mpirica l results [3] for 16 c lie nts to a muc h larger network, with rea listic c lie nt assoc iation

   interfere nce , Re-configuration of APs a nd channe ls with fee dba c k infonna tio n.

   V Shrivasta va in his work [6] captures a signific ant research and engi ne ering effort in e xploring the role of centra lization in enterprise WLANs and ma kes the following c ontributions:

   • Demonstrates the importanc e of a ddressing downJink hidden and e xpose d te nnina l proble ms: We start by de monstrating that we are solving a practica l proble m that occurs in e nterprise \VLAN settings. We show that downlink hidden and exposed terminals are preva le nt in multiple enterprise \VL ANs through analysis and

     measurement of production WLANs, as well as measurements on our testbeds. We qua ntify the performa nce loss observed due to hidden and exposed termina ls in such settings.

     • Demonstrates the role of selective data-path ce ntralization in enterprise WLANs and how it can be impleme nte d independe ntly by a single enterprise WLAN vendor: We show that a selective amount of data-path ce ntralization is useful in enterprise WLANs in directly mitigating perfonnance loss due to downlink hidden and exposed terminal scenarios. Further, such a mechanism can indirectly he lp improve the performance of the entire WLAN environment. All proposed mechanisms require no changes in clie nts and hence can be imple me nte d sole ly by an enterprise \VLAN vendor.

     • Implements and deploys CENTAUR over two different testbeds and platforms: We imple ment CE l\'rfAUR over two different testbe ds, each with a different wireless pla tform, NlC, and wired bac kpla ne . (i) Testbed l: locate d across five floors of a building consisting of 30 266-MHz Soekris 4826 nodes equippe d with Atheros-based 802.11 NICs deployed and interc onnecte d with a 100 Mbps Ethernet backplane, and (ii) Testbed 2 deployed across a single floor consisting of 20 1.2-GHz VIA nodes equipped with Intel 2915 802.11 ABG NICs deployed in a single floor of a building and interconnecte d with a Gigabit Ethernet backplane.

   Evaluates CENTAUR usmg controlle d experiments and playback of real traffic traces: We evaluate the performance of CE NT AUR through a combination of controlle d experiments as well as by playing back real traffic traces on these testbeds. We use different metrics for all our measurements including throughput (UDP and TCP), fairness, completion time of web transactions (http downloads), and MOS for VoIP-like traffic . Example results from our experiments on pla yback of real traffic traces, under observed periods of high loads, and averaged over all traffic across an enterprise WLAN, include: up to 1.48 x improvements in data throughp uts, 1.38 x reductions in web transaction completion times, and 1.21 x improveme nts in MOS for VoIP-like traffic.

   for individua l hidden and exposed termina l links are obviously much higher.

   J Manweiler in his work [7] motivating MIM-aware revisions to link-scheduling protocols. He identifies the opportunity m MIM-aware reordering, characterizes the optimal improve ment in throughput, and designs a link-la yer protocol for enterprise wireless LANs to ac hie ve it. Testbed and simulation results confirm the performa nce gains of the proposed system.

   M. Lacage in his work [8] focus on the task of ma.ximizing the application-leve l throughput in infrastructure networks through practical rate· adaptation algorithms. Because no published pa per discusses the issues surrounding real imple me ntations of 802.11 rate adapta tion algorithms, we believe our ma in contribution to be the identification of two classes of 802.11 devices: low late ncy and high la te nc y systems. Low late nc y systems allow the imple menta tion of per-packet adaptation algorithms while high latenc y syste ms require periodic analysis of the transmission characte ristics and updates to the transmission parameters.

   H. Falaki in his work [9] study the interaction of smartphone traffic with the adio power management polic y. They find that the current slee p timers, that is, the idle period after which the radio will go to sleep, are overly long. By reducing them based on current traffic patterns, radio power consumption can be reduced by at least 35% with minimal impact on performa nce.

   Smartphone traffic represents an increasingly large share of Internet traffic. Ce llular traffic is projected to grow IO times faster than fixed Internet traffic [!OJ and most of this traffic is generated by smartphones [I I]. By next year, smartphone sa les are projected to smpass desktop PCs [12]. However, little is known today about the nature of smartphone traffic. Two recent studies have shed valuable light on some aspects of this traffic.

   Trestianet a l. study the kinds of Web sites accessed at diff erent times of the day [13). Maier et al. study HTTP traffic generated by mobile handheld de vices (which include music players and personal gaming consoles in addit ion to smartphones) in homes [ 14]. Both studies are based on data gathered at a link in the middle of the network. As a result, while they can analyze traffic from a large number of devices, they do not capture a detailed, comprehe nsive view of individual devices. For instance, the second study misses traffic exchanged by devices through the cellular interf aces or outside of their homes.

   Arpit Gupta in his work [15] demonstrates that solving the problem of tr affic asymmetry results in maximum perf ormance impr ovements for large audience envirorunents. They find corre lation between the presence of asymmetry in network traffic and instantaneous transmiss ion queue al the WiFi AP and develop a mechanism where traffic asymmetr y is inferred in real time, pr ior itizing the AP accordingly for channel resource access over competing ST As. For large audience envir onments, the pr ior itization of AP's tr affic enables effic ient realization of AP-only f airness solutions. The ke y contribution of our work is the empirical study of the performanc e implications of these solutions in order to optimize the perf ormance of busy WiFi hotspots. To add realism to our results, we implemented our solutions in an off-the-shelf

   commercial IEEE 802.llg AP, constructed a real network testbed of 45 WiFi nodes and tested the

   graph to accurately model the RF envir onment without making path loss assumptions. Utility functions are defined on the conflic t graph to characterize network performance. Fina lly, a variety of operating parameters are used to optimize the computed util ity.

   J. Elson in his work [ 17] presents their idea for post-f acto synchronization, an extremely low-power method of synchronizing cloc ks in a local are a when accurate timestamps are needed for spec if ic eve nts . We also present an experiment that suggests this multi-modal scheme is capable of precision on the order of 1 msec-an order of magnitude better than either of the two modes of which it is composed. These results are encouraging, although still preliminary and performed under idealize d laboratory condit ions.

   B. A. Mah in his work [18) developed an empirical

   model of network traffic produced by HTTP. Instead of relying on server or client logs. our approach is based on packet traces of HTTP conversations . Through traffic analysis, author have determined statistics and distributions for higher leve l quantities such as the size of HTTP files, the number of files per "web page", and user browsing behavior. These quantities form a model can the n be used by simulations to mimic World Wide We b network applications.

   Y. Bejerano in his work [19) present the Mana ged WiFi sys tem, called MiFi, f or supporting fairness

   -performance of various o timization setting-s in and Qo. S in the existing- IEEE 802.11 MAC. layer.

   network traffic loads emulating the traffic patterns captured from real traces.

   N. Ahmed in his work (16] present SMART A, an architecture that takes the above-mentioned challe nges into account. Our infrastr ucture based solution, targeted towards enterprise wireless LANs, does not require client-s ide modif ic ations, allowing backwards compatibil it y. Utility functions provide a unif ied framework for capturing multiple and even conflicting performance objectives. Mor eover , SMARTA makes no assumptions about RF propagation and uses dynamic optimization to address varying channel condit ions. At a high level, SMARTA uses active probes to build a con- flict

   To the best of our knowledge this is the first comprehens ive system that overcomes both the hidden node and the over lapping cell problems in multiple-AP \VLAN netw orks.

   C Coutras in bis work (20] propsed the 802.11 MAC layer protocol provides as ync hronous, time bounde d, and contention free access control on a variety of physical layers . In this paper we examine the ability of the point coordination function to support time bounded services. We present our proposal to support real- time services within the framework of the point coordination function an d discuss the spec if ics of the connec tion establishment pr ocedure.

2. CONCLUS ION AND E N H A N C E M E N T S

The paper summarizes the current works in the channel scheduling in WLAN to reduce the

1. N. Wood. Mobile NdaCtEa TtrEaIfTfic- g20ro17wCthon10ferteimncees Pfraoscteeerdtihnagns fixed over next five years. http://www.totaltele .co m/vie w.aspx?ID=448681.

2. Mobile data traffic surpasses voice. http://www.e ricsson.com/thecompany/press/releases/2010/03/

interference. We explored the problems in the 1396928.

current solutions and identified the openi areas for concrete solutions for these open problems to enable effective channel utilization by reducing the interference.

R EF E R E N C ES

1. A. Patro, S. Gov indan, and S. Banerjee , Observing Ho me Wireless Experience Through WiFi APs, inProceedings of ACM MobiCo m, 2013.

2. M. A. Ergin, K. Ra machandran, and M. Gruteser, Understanding the effect of access point density on wire less lan performance, in Proceedings of ACM MobiCo m, 2007.

3. S. Choi, K. Park, et a l. On the performance characteristics of W LANs: rev isited. SIGM ETRICS Perfo rm. Eva l. Rev., 33(1):97 108, 2005.

4. R. Gu mmadi, D. Wetherall, B. Greenstein, and S. Seshan, Understanding and mitigating the impact of rf interference on

   802.11 networks, in Proceedings of ACM SIGCOMM, 2007.

5. Y. Lee, K. Kim, and Y. Choi, Optimizat ion of ap place ment and channel assignment in wire less lans, in Proceedings of IEEE LCN, 2002.

6. V. Shrivastava, N. Ahmed, S. Rayanchu, S. Banerjee, S. Keshav, K. Papagiannaki, and A. Mishra, CENTAUR: Realizing the Full Potential of Centra lized Wlans Through a Hybrid Data Path, inProceedings of ACM MobiCo m, 2009.

7. J. Manwe iler, N. Santhapuri, S. Sen, R. Choudhury, S. Nela kuditi, and K. Munagala, Order matters: Transmission reordering in wire less networks,IEEE/A CM Transactions on Networking, vol. 20, no. 2, pp. 353366, Apr. 2012.

8. M. Lacage, M. H. Manshaei, and T. Turletti, IEEE 802.11 Rate Adaptation: A Practica l Approach, inProceed ings of ACM MSWiM, 2004.

9. H. Fala ki, D. Ly mbe ropoulos, R. Mahajan, S. Kandula, and D. Estrin, A First Look at Traffic on Smartphones, inProceedings of ACM IMC, 2010.

1. L. Snol. More s martphones than desktop PCs by 2011. http://www.pcwo rld.co m/a rtic le/171380/mo re_smartphones_t han_desktop_pcs_by_ 2011.ht ml.

2. I. Trestian, S. Ranjan, A. Kuzmanovic, and A. Nucci. Measuring serendipity: Connecting people, locations and interests in a mobile 3G network. In IM C, 2009.

3. G. Ma ier, F. Schneider, and A. Feldmann. A First Look at Mobile Hand-He ld Dev ice Tra ffic. In PAM, 2010.

4. A. Gupta, J. Min, and I. Rhee, Wifo x: Sca ling wifi performance for la rge audience environments, in Proceedings of ACMCoNEXT, 2012.

5. N. Ah med and S. Keshav, Sma rta: A self-managing architecture for thin access points, in Proceedings of ACM CoNEXT, 2006.

6. J. Elson and D. Estrin, Time synchronization for wireless sensor networks, in Proceedings of the 15th International Sy mposium on Paralle l and Distributed Processing, Apr. 2001, pp. 19651970.

7. B. A. Mah, An e mp irica l model o f http network traffic, inProceedings of IEEE INFOCOM, 1997.

8. Y. Be jerano and R.S. Bhatia. M iFi: a fra me work for fairness and QoS assurance for current IEEE 802.11 networks with mu ltip le access points. IEEE/A CM Transactions on Networking, 14:849 862, aug. 2006.

9. C. Coutras, S. Gupta, and N. B. Shro ff. Scheduling of real-t ime t raffic in ieee 802.11 wireless lans. Wire less Networks, 6(6):457 466, 2000.