Priv-Code: Preserving Privacy Against Traffic Analysis through Network Coding for Multihop Wireless Networks
|
|
- Lydia Kennedy
- 8 years ago
- Views:
Transcription
1 Prv-Code: Preservng Prvacy Aganst Traffc Analyss through Networ Codng for Multhop Wreless Networs Zhguo Wan, Ka Xng, Yunhao Lu MOE Key Lab for Informaton System Securty, School of Software, Tsnghua Natonal Lab for Informaton Scence and Technology, Tsnghua Unversty, Department of Computer Scence and Engneerng, HKUST, School of Computer Scence and Technology Unversty of Scence and Technology, Chna, Abstract Traffc analyss presents a serous threat to wreless networ prvacy due to the open nature of wreless medum. Tradtonal solutons are manly based on the mx mechansm proposed by Davd Chaum, but the man drawbac s ts low networ performance due to mxng and cryptographc operatons. We propose a novel prvacy preservng scheme based on networ codng called Prv-Code to counter aganst traffc analyss attacs for wreless communcatons. Prv-Code s able to provde strong prvacy protecton for wreless networs as the mx system because of ts ntrnsc mxng feature, and moreover, t can acheve better networ performance owng to the advantage of networ codng. We frst construct a hypergraphbased networ codng model for wreless networs, under whch we formalze an optmzaton problem whose objectve functon s to mae each node have dentcal transmsson rate. Then we provde a decentralzed algorthm for ths optmzaton problem. After that we develop an nformaton theoretc metrc for prvacy measurement usng entropy, and based on ths metrc we demonstrate that Prv-Code acheves stronger prvacy protecton than the mx system whle achevng better networ performance. I. INTRODUCTION Traffc analyss s a powerful tool to deduce nformaton from communcaton patterns, no matter whether the messages are encrypted or not. Numerous traffc analyss attacs have been successfully appled to varous communcaton networs, ncludng both mltary and cvlan systems. Due to the open nature of wreless medum, wreless networs are specally vulnerable to traffc analyss attacs. Wreless telegraphy and rado networ are two typcal examples attaced by traffc analyss. To fght aganst traffc analyss attacs, Davd Chaum [1] proposed the dea of mx to hde correspondence between message senders and recevers, and hence to protect communcaton prvacy. Specfcally, a Mx accepts a number of fxedlength messages from sources, performs cryptographc transformatons on the messages, and then forwards the messages to the next destnaton n an order unpredctable from the order of nputs. An obvous feature of the mx-based system s that messages are cached and reordered at each Mx before beng sent out, and the content of each message s not changed after the mxng process. Followng the dea of Chaum, a number of anonymous communcaton systems have been proposed, ncludng Crowds [2] and Tor [3] for web browsng, Mxmaster [4], and Mxmnon [5] desgned for emal prvacy. Although the mx-based mechansm can be appled to protect prvacy, ts defcences are obvous. Frst of all, each Mx has to cache enough messages before sendng them n a random order, whch ntroduces unpredctable delay nto the system. As a result, the mx-based system lacs ablty to support realtme networ traffc or guarantee certan qualty of servce. Next, the mx-based system requres each Mx to perform decrypton and re-encrypton, normally publc ey based operatons, for each message. Ths leads to too much computaton cost for Mxes, and also ncreases message transmsson delay. Last but not least, efforts n ths lne of research have been focused on prvacy preservaton usng cryptographc prmtves, wthout consderng networ performance. Therefore t s mperatve to desgn an anonymous communcaton system wth the performance concern n the prvacy preservng desgn, as the ncreasng popularty of resource-constraned wreless devces. In ths paper, we tacle the prvacy-preservng communcaton system desgn problem based on the networ codng technque. The concept of networ codng was frst proposed by Ahlswede et al. n 2000 [6], and has been vewed as a promsng technology for mprovng networ performance and enhancng networ relablty. The networ codng mechansm wors dfferently from tradtonal routers or Mxes n that messages are coded at ntermedate nodes, n contrast to message reorderng n mx-based systems. Although networ codng s proposed as a tool to mprove networ performance and relablty, t has the ntrnsc mxng feature le the Mx n mx-based systems. Intutvely, Mxes n mx-based systems can be modfed to encode messages nstead of smple reorderng for better performance and relablty. However, there has been very lmted research on employng networ codng to counter aganst traffc analyss, and most wor on traffc analyss focuses on mx-based systems. Untl recently, the potental of networ codng on reslence to traffc analyss has been noted by Fan et al. [7]. They
2 2 analyzed prvacy enhancement wth networ codng n case of traffc correlaton attacs, whle formal treatment of prvacy mprovement due to networ codng schedulng s not gven n ther paper. And ther scheme s desgned for wred networs, wthout consderng the broadcast nature of wreless medum. In ths paper, we propose Prv-Code, a networ codngbased scheme to preserve prvacy aganst traffc analyss for wreless communcatons. Ths scheme formalzes the prvacy preservng problem as an optmzaton problem under a hypergraph-based networ codng model for wreless networs. It provdes a decentralzed algorthm for ths optmzaton problem. We develop an nformaton theoretc metrc for prvacy measurement usng entropy, and based on ths metrc we demonstrate that Prv-Code acheves stronger prvacy protecton than the mx system whle achevng better networ performance. The contrbutons of our wor can be summarzed as follows: We defne a networ model usng the drected hypergraph smlar to [8] for networ codng schedulng n multhop wreless networs. The model not only captures the broadcast nature of wreless networs, but also consders the lossy characterstc and the MAC nterference of wreless medum. We formalze the prvacy-preservng networ codng schedulng problem as an optmzaton problem, and provde a decentralzed algorthm based on decomposton technques. An nformaton theoretc metrc for quanttatve prvacy measurement based on nformaton entropy s proposed n ths paper. It provdes a general way to evaluate prvacy protecton strength of varous mechansms wth regard to traffc flow nformaton. We mplement the algorthm and conduct experments for dfferent networ parameters to evaluate ts prvacy protecton and performance. The proposed scheme not only provdes strong prvacy protecton but also has good networ performance. The rest of the paper s organzed as follows. In the next secton we revew related wor on the mx system and networ codng. Then our scheme on explotng networ codng to acheve prvacy s descrbed n detal n Secton III. In Secton IV we analyze and dscuss ssues of the proposed scheme on prvacy protecton. Detals on the smulaton to evaluate the proposed scheme are provded n Secton V. Secton VI summarzes and concludes the paper. II. RELATED WORK A. The Mx-based Systems Snce Chaum s mx system was proposed, many smlar desgns have been ntroduced n the lterature [2], [3], [4], [5], [9]. In these systems, a mx-net s formed by a set of mx nodes, and messages are mxed when they traverse the mx-net to acheve anonymty. Generally, these systems can be grouped nto real-tme systems and non-realtme systems. Crowds [2] and Tor [3] belong to the real-tme mx systems. In Crowds, a user jons a Crowd as a jondo, and each jondo acts as a proxy passng web requests to a random Crowd member or the web server accordng to a gven probablty. Thus, Crowds s able to preserve anonymty aganst collaboratng members, but only recever anonymty s provded n case of local eavesdroppng. Cypherpun, Mxmaster [4], and Mxmnon [5] are non-realtme systems beng used as anonymous re-malers. Cypherpun s the frst wdely mplemented mx-le system, wthout features le the message paddng or message pools, whch maes t the Type I anonymous remaler. Mxmaster, the Type II anonymous remaler, fxed these problems n Cypherpun, whle the Type III remaler Mxmnon further enhance prvacy by mang reply and forward messages ndstngushable. As analyzed, the man drawbacs of the mx-based systems are ts low networ performance and hgh computaton cost. Informaton theoretc treatment on prvacy metrc n mxbased systems s studed n [10]. The anonymty s measured by nformaton entropy of the mx-based systems, whch s able to accurately characterze prvacy protecton strength. The anonymty entropy measures the uncertanty that the attacer on dentfyng a sender or recever. Let the probablty dstrbuton of a user u beng the sender (or the recever) of a message M be p u, where p u =1, then the anonymty u metrc can be computed as H(M) = p u log p u. Ths u entropy represents the anonymty status of the message M, and the number of bts of addtonal nformaton the attacer needs to dentfy the message sender. In ths paper, we use a smlar nformaton theoretc method to analyze prvacy protecton acheved for wreless networs wth our scheme. B. Wreless Networ Codng The networ codng technque [6] was orgnally proposed as a soluton to mprove networ performance and soon receved extensve attenton n the networng communty. Ths technque has been extensvely studed n wreless envronments to fully explot the broadcast nature of wreless medum. Katt et al. [11] have mplemented a real networ codng system COPE that XOR pacets n the wreless networ. The performance mprovement of COPE can be up to 70% for wreless mesh networ, and COPE even has 3-4 tmes of throughput gan on the testbed. Research on wreless networ codng has studed networ codng schedulng for dfferent purposes or applcatons, rangng from broadcast, multcast, relablty, energy effcency, to maxmzng throughput. Securty and prvacy n networ codng have also been mportant research drectons. There have been a lot of research on securty ssues n networ codng, e.g., [12], [13], [14], [15], [16], [17], [18]. However, the prvacy ssue of networ codng has been largely gnored by the research communty. Untl recently, a scheme proposed by Fan et al. [7] s the only one to enhance prvacy by networ codng. They employed networ codng to counter aganst traffc analyss attacs n wred networs, and use homomorphc encrypton to protect
3 3 code coeffcents of messages. Although they analyzed prvacy enhancement due to networ codng n case of pacet sze correlaton, tme correlaton and content correlaton, formal treatment of prvacy wth regard to traffc flows s not gven n ther paper. That ts, they dd not answer the queston on how to schedule networ flows, the central part of networ codng, n order to mprove prvacy. Also the proposed soluton s desgned only for wred networs, and cannot be used n wreless networs. Furthermore, the advantages of networ codng on performance mprovement and energy savng are not fully exploted for ether wred networs or wreless networs. III. THE PROPOSED SCHEME: PRIV-CODE The proposed scheme Prv-Code consders concurrent u- ncast sessons n mult-hop wreless networs, and employs ntra-sesson networ codng for data communcaton,.e., only pacets from the same sesson can be encoded together. The man goal of Prv-Code s to acheve strong prvacy aganst traffc analyss wth proper networ codng schedulng. The ntutve behnd Prv-Code s that one can mae all nodes n the networ transmt traffc wth the same traffc pattern, then the attacer s not able to dstngush traffc senders or recevers. If all nodes n the networ transmt data wth the same data rate, and all transmsson flows loo no dfferent from each other for varous traffc analyss attacs, then the attacer cannot obtan any nformaton on senders or recevers at all. Ths can be done by usng specally desgned networ codng schedulng whch schedules an end-to-end uncast sesson over multple paths. It wll not only sgnfcantly enhances resstance aganst traffc analyss, but also mproves networ performance due to benefts of networ codng. In ths secton, we frst ntroduce assumptons and threat model n our scheme, then we descrbe our system model whch captures the broadcast nature and lossy characterstc of wreless medum. Next, we propose an optmzaton framewor to fnd the optmal networ codng schedulng for enhanced prvacy. After that, we provde a decentralzed algorthm for ths optmzaton framewor, whch s specally desgned for dstrbuted multhop wreless networs. A. Assumptons and Threat Model In ths paper, we assume a mult-hop wreless networ, n whch each node s equpped wth only one antenna. All nodes n the networ have dentcal transmsson range as well as nterference range. We also assume that the wreless medum s lossy. An anonymous routng protocol for mult-hop wreless networs provdng anonymty and unobservablty s mplemented as [19]. Hence routes can be securely and anonymously establshed from a source to some destnatons, and an outsde attacer cannot access the pacet header to now the pacet type, or source/destnaton address. In order to avod pacet sze correlaton, all pacets are of the same sze. We assume exstence of a global adversary who can passvely montor the whole networ. He can contnuously observe the entre wreless networ, and hence obtan traffc flow nformaton ncludng node transmsson rates, nter-pacet ntervals etc. The attacer can mae use of exstng traffc analyss technques used n [20] and [21]. However, he cannot decrypt any encrypted pacet wth brute force attac. Hs goal s to deduce who s the sender or recever of a message from networ traffc nformaton. B. The System Model The data flow of a sesson s dvded nto generatons, and pacets from the same generaton can be encoded. The number of pacets n n a generaton can be confgured to sut the applcaton. At each node, random lnear networ codng s used to process the pacets, and the encoded pacets are transmtted to the destnaton va multple paths. How to establsh multple paths wll be descrbed n the next secton. The networ s modeled as a drected hypergraph H =(N, A) as n [8], where N s the set of nodes and A s the set of hyperarcs. A hyperarc (, J) s formed by a start node and a set of end nodes J, whch s a non-empty subset of N. Each hyperarc (, J) represents a broadcast ln from node to nodes n J. As we assume all nodes have dentcal transmsson range, each node has a unque end node set J. Ths defnton captures the broadcast nature of wreless medum. The hypergraph s degraded nto a conventonal graph model when J contans only one node. A set of uncast sessons U = {u 1,..., u U } s transmtted through the networ. Let S and T ( =1, 2,..., U ) be the source and recever of the uncast sesson, and r denote the flow rate of sesson. For a uncast sesson where the source S wants to send data wth rate r to T, by the flow conservaton condton, we have: fjj fji = δ, N, (1) j N {I (j,i) A, I} where r, f = S δ = r, f = T 0, otherwse and fjj s the flow rate over hyperarc (, J) ntended to node j J. For sesson, the equaton represents the flow conservaton constrant that the source node s net transmsson rate s r, the destnaton node s net transmsson rate s r, and any ntermedate node s net transmsson rate s 0. Under the hypergraph model, we further set up the broadcast MAC model to characterze the nterference n wreless networs, and the codng model how pacets are coded wth networ codng. We use the broadcast MAC model of Zhang and L [22], whch extends the uncast MAC model to obtan a necessary condton for feasble broadcast schedules. In ths model, the transmsson range and the nterference range are consdered to be the same, and the recepton probablty beyond ths range can be gnored. Specfcally, the wreless networ s modeled as an deal tme-slotted broadcast MAC where competng transmtters can
4 4 optmally multplex the channel wthout any collsons. For a uncast sesson, let B [t] (0 or 1) be the decson varable ndcatng whether node s transmttng n slot t, and I() be the set of all transmtters wthn s range (ncludng ). Under the hypergraph model, I() s equal to node s end node set J. Then a schedule s collson free ff: B [t]+ Bj [t] 1, N\S. (2) j I() Ths equaton ndcates that any recever allows the broadcast transmsson from at most one transmtter wthn ts range at each tme slot. Denote T as the perod of a schedule, and b as the rate at whch node broadcasts pacets to ts downstream nodes, then we have: b 1 = lm T T T B [t]. (3) t=1 Apply (3) to (2), we can have: b + b j C, N\S. (4) j I() where C = 1 T s the MAC layer capacty, equalng to the maxmal broadcast rate of a node when no nterference presents. Meanwhle, t s noted that a constrant on capacty for the broadcast ln (, J) should be satsfed. Snce we assume lossy wreless lns n our scheme, the recepton probablty of the ln on the hyperarc (, J) to node j s p Jj. Then we have the followng ln capacty constrant: b p Jj f Jj. (5) C. Schedulng Optmzaton for Prvacy Before presentng the schedulng optmzaton algorthm for prvacy, we frst show how to mae a traffc flow ndstngushable from another flow. The queston s what nd of traffc pattern should the data flows tae so as to mae them ndstngushable. The traffc pattern of a flow s determned by pacet arrval rate, nter-pacet ntervals and arrval tme dstrbuton. If message arrvals of traffc flows are Posson processes wth the same arrval rate, then the correlaton attac wll be neffectve. Consequently, t s suffcent and necessary to shape the traffc flows as Posson processes wth the same arrval rate n order to counter aganst the traffc analyss attacs based on these propertes. In ths case, the nter-pacet ntervals are exponentally dstrbuted and pacet arrval tmes are unformly dstrbuted. More dscusson s left n Secton IV. Based on the above concluson, what we need to do s to mae every traffc flow have the same rate and shape the traffc flows to be Posson processes wth the same arrval rate,.e., the same flow rate. As we assume an unobservable routng protocol s used, the flow rate that the adversary can observe s cumulated across concurrent uncast sessons. Meanwhle, t s desrable to mae the transmsson rate of as low as possble for energy effcency. Thus, we formulate the followng schedulng optmzaton problem: Prvacy-Mnmax: mn b,f max R (6) subject to: fjj fji = δ,,, j N {I (j,i) A, I} b = R b + b j C, S, j I() fjj b p Jj,, j J,. (7) where N, Kwth K beng the set of all concurrent sessons n the networ, and r, f = S δ = r, f = T 0, otherwse. In ths optmzaton problem, the objectve s to mnmze the maxmum flow rate R for all nodes. Though t would not acheve the maxmal prvacy entropy as nodes flow rate R s may be dfferent, the dfference between R and R j for two nodes and j s reduced. We can then nject paddng traffc nto each node to mae R be the same. A more mportant result s ths objectve functon tres to eep nodes transmsson rates low globally. The problem can be equvalently formulated as follows: Prvacy-Mnmax : mn b,f,t R (8) subject to: fjj fji = δ,,, j N {I (j,i) A, I} b + b j C, S, j I() fjj b p Jj,, j J,, b R,. (9) After a soluton of the schedulng optmzaton problem s obtaned, all nodes may have the same transmsson rate n the deal case. Then each node can transmt data n accordance to Posson dstrbuton, so that the adversary s not able to dstngush them. For the cases where not all nodes have the same transmsson rate (.e. R = b ), we can smply nject dummy traffc at nodes to reach the maxmum transmsson rate R so they have the same transmsson rate. D. A Decentralzed Networ Schedulng Soluton Though the Prvacy-Mnmax problem can be readly solved by standard lnear programmng algorthm, t s desrable to provde a decentralzed soluton for the networ schedulng problem. In ths secton, we propose a decentralzed algorthm for the networ schedulng problem based on decomposton technques [23]. Specfcally, we decompose the orgnal problem nto three separate subproblems wth decoupled varables based on the dual decomposton. Then
5 5 we solve the subproblems ndependently, and fnally solve the master dual problem by updatng dual varables. We frst ntroduce Lagrange multplers λ, μ, νj, and ϕ to relax the four sets of constrants n (9) respectvely. Then the Lagrangan functon s as follows: L(b, f, R, λ, μ, ν, ϕ) =R + λ ( fjj fji δ ) j N I μ ( b + b j C) j I() νj(f Jj b p Jj ) j I() ϕ ( b R) Note that n our model the nterference node set I() s equal to node s end node set J. Thus, the orgnal problem can be decomposed nto three ndependent subproblems as follows: SUB2: F 2 = mn f SUB3: F 3 = mn b SUB1: F 1 = mn (1 R ϕ )R (10) fjj(λ λ j + νj) (11) b (μ + (μ j νjp Jj)+ϕ ) (12) j I() Snce all the three subproblems are lnear, the Lagrange multpler method may not necessarly generate the optmal soluton. We adopt the proxmal method and add a quadratc term to mae t strctly convex [24]. Tae SUB2 as an example, the optmzaton problem to be solved at each node s: mn f fjj(λ λ j + νj) Then we can add a quadratc term nto t whch turns the optmzaton problem to be: mn f f Jj(λ λ j + νj)+ 1 2c f Jj fjj(t) 2 Then fjj s updated by fjj(t +1)=[fJj(t) c(λ λ j + νj)] + where c s postve constant scalar that maes the above update to be arbtrarly close to the optmal value of fjj, and [ ]+ denotes the projecton onto the non-negatve orthant. After the above three subproblems have been solved by each node for gven λ, μ, ν, ϕ, we proceeds to solve the followng master dual problem. max F 1 + F 2 + F 3 + λ δ + μ C (13) λ,μ,ν,ϕ subject to: μ > 0,ν j > 0,ϕ > 0 (14) where F 1, F 2, F 3 are solutons to the subproblems for gven λ, μ, ν, ϕ. The subgradent method can be used here to fnd the optmal soluton to the master dual problem. In each teraton of subgradent optmzaton procedure, each Lagrange multpler s updated accordng to ts subgradent. For nstance, λ s updated n each teraton by: λ (t+1) = [λ (t)+α(t)( f Jj f ji δ )] + (15) j N I where t s the ndex of the teraton, f Jj s the optmal soluton from subproblem SUB2, and [ ] + denotes the projecton onto the feasble set of λ. α(t) s the step sze for teraton t. A dmnshng step sze s adopted for the purpose of convergence. Specfcally, α(t) = A 1+B t where A and B are non-negatve tunable system parameters. To summarze all the above procedures, we formulate the followng decentralzed Algorthm 1: Algorthm 1 The Decentralzed Prvacy-Orented Schedulng Optmzaton Algorthm Input: Hypergraph: (N, A), Sesson Set: K, Flow rate set: R, Ln recepton probablty set: P, Ln Capacty: C. Output: fjj and b. 1) Intalzaton: set t =0, λ equal to some ntal value, and μ,νj,ϕ equal to some non-negatve values for all A,j J, K, where (, J) N. 2) Each node locally solves ts subproblems from SUB1, SUB2 and SUB3 for each sesson K, and then broadcast the result fjj and b to ts drect neghbors,.e., all nodes j where j J for (, J) A. 3) Each node updates the Lagrange multplers by the subgradent method as llustrated n (15). Then t broadcast ϕ (t+1) to all other nodes, and broadcast λ (t+1),μ (t+1) to ts drect neghbors. 4) Set t t+1 and go to step 2 (untl satsfyng termnaton crteron). It s mportant to note that t s unnecessary to broadcast νj (t +1) to other nodes, and λ (t +1),μ (t +1) need to be broadcast to s neghbors only. The convergence of the above algorthm follows the general convergence propertes of subgradent and dual decomposton method. After the optmal soluton for the objectve functon s obtaned, each node stores, encodes, and forwards pacets towards the next hop wth calculated transmsson rate b at each sesson. As the resultng networ schedulng satsfes the flow constrants and capacty constrants, we conclude that the schedulng can acheve expected data rates. In order to mae each node have dentcal traffc pattern, we requre each node to add approprate dummy traffc. An obvous soluton s to nject dummy traffc to the maxmum transmsson rate R. A good way to create dummy traffc s create redundant or lnearly dependent pacets, as these pacets can mprove relablty n lossy wreless transmsson.
6 6 Then each node transmts pacets n accordance to Posson dstrbuton. IV. TRAFFIC ANALYSIS AND PRIVACY METRIC In ths secton, we frst demonstrate that traffc analyss attacs explotng traffc patterns on nter-pacet ntervals, pacet arrval tme or data rates are neffectve n dstngushng data flows conformng to Posson dstrbuton wth dentcal arrval rate. We do not consder content correlaton attacs as n [7] snce we assume that no content correlaton nformaton s leaed. Based on ths result, we present an nformaton theoretc metrc on prvacy for wreless networs. Then we use t to measure prvacy qualty offered by our scheme Prv-Code. As assumed, there s an anonymous routng protocol whch can provde unobservablty [19] mplemented n the wreless networ. The adversary s not able to now the content of any pacet, ncludng the address feld n the pacet header. A. Traffc Analyss A node wth an ncomng flow conformng to Posson dstrbuton and exponental delay tmes can be vewed as an M/M/1 queung system. Accordng to Bure s theorem, the departure process of an M/M/1 queue s also a Posson process wth the same rate ndependent of the arrval process. Ths feature helps the proposed scheme Prv-Code thwart traffc analyss attacs. If ncomng flows and outgong flows are ndependent Posson processes wth the same arrval rate, t s mpossble for an attacer to dstngush them. Traffc analyss attacs have been proposed to explot dfferent traffc patterns of a flow, e.g., pacet delay characterstc [20], number of pacets n a fxed nterval or wndow [21], [25]. Danezs [20] uses maxmum lelhood estmaton to dstngush dfferent flows based on pacet delay characterstcs. Pacets are delayed for a perod conformng to the exponental dstrbuton at the mxes, whch s the optmal mxng strategy for a contnuous-tme mx networ. Let X and Y are two output lns of an exponental mx that the attacer wants to dfferentate, C X and C Y are two model probablty dstrbutons for the two output lns. Then the lelhood rato can be formulated as n C X (X ) m u L = =1 j=1 n u m, C Y (Y j ) =1 j=1 where u s the unform dstrbuton parameter, and X and Y j are sampled tmes comng out of channel X and Y respectvely. It can be verfed that f C X and C Y are unform dstrbutons then the lelhood rato s 1, whch s the case when the ncomng traffc to the mx s a Posson process. So we can see the attac fals when all traffc flows follow Posson dstrbuton wth the same arrval rate. Zhu et al. s approach [21] correlates flows usng number of pacets n n a fxed nterval. The pattern vector X of an nput ln or an output ln n Zhu et al. s approach contans the followng element: Number of pacets n batch X, =. Tme elapses n batch Then the mutual nformaton between X and another pattern vector Y j s I(X, Y )= p(x,y )log p(x,y ) p(x )p(y ). Ths attac s especally effectve aganst TCP due to the TCP loop-control mechansm. However, f the nput flow and output flow conform to ndependent Posson dstrbuton as n our scheme, the attac wll fal snce the mutual nformaton s actually 0. A tmng analyss attac proposed n [25] adopts a smlar strategy as [21]. Specfcally, for each possble entry-ext par, the attacer computes the cross-correlaton of the two sequences as r(d) = Σ ((x μ)(x μ )) Σ (x μ) 2 Σ (x μ ) 2, where x s the number of pacets receved or sent by a mx durng the th wndow, μ and μ are means of the two sequences. Correlaton on the nter-pacet ntervals between two networ lns may lead to concluson that they are carryng the same traffc. However, the cross-correlaton s 0 for two ndependent Posson processes, so the attac wll not succeed n Prv-Code. The adversary can launch two types of actve attacs: artfcal gaps and artfcal bursts [25]. In the artfcal gap attac, the adversary gets control of some vald nodes n the wreless networ, and selectvely drops several consecutve pacets n a target flow to create a gap, whch wll result n a gap n other lns. By examnng the change on the nterpacet nterval pattern, the adversary can dentfy related lns. For artfcal gaps, the attac must be prevented by njectng dummy traffc nto nodes beng nfluenced, so as to mae the traffc pattern unchanged. The artfcal burst attac s to create a traffc burst by holdng up pacets at some nodes and release them at once. But such attacs can be effectvely thwarted at each node by modulatng outgong data as a Posson process for a predetermned rate. Moreover, the burst pacets may be relayed through multple paths, thus the burst s releved by at each of the multple lns. As a result, the rs of one output ln beng found to be related to an nput ln s removed. B. Informaton Theoretc Metrc for Prvacy We adopt and extend the nformaton theoretc approach for prvacy measurement n [10], whch proposed a prvacy metrc for mx networs. Ths metrc s desgned for mx networs, whch have several dfferences wth the networ model used n our scheme. The man dfference s that a node n our networ model can be a sender or a recever when t wors as a mxng relay for others, whle a mx n a typcal mx networ
7 7 s normally not a sender or a recever. Another dfference s that nodes n n our networ model may form a loop, whch does not exst n the mx networ. Hence, we have to adapt the nformaton theoretc metrc for prvacy to our networ model. The prvacy metrc defned n [10] uses entropy to descrbe prvacy qualty. Specfcally, sender anonymty (or recever anonymty) s the entropy of the attacer s probablty dstrbuton of users beng the sender (or recever) wth respect to a message. Let Ψ be the set of all users, u Ψ be the user, and p u be the probablty of the user u beng the sender of a message M. Then the prvacy measurement, called the effectve anonymty sze, of the sender anonymty wth respect to M s: H(M) = u Ψ p u log p u. The prvacy metrc can be nterpreted as the number of bts of addtonal nformaton that the attacer needs to dentfy the user u beng the message sender. It s trvally to see that f p u =1for some user u then the entropy s 0 bts, meanng the attacer has dentfed the user already Fg. 1. A Smple Mx Networ and Its Prvacy Metrc. The sender anonymty of the output lns of mx 1 s H1 total = ( 1 2 log log 1 )=1, and 2 the sender anonymty of the output lns of mx 3 s H3 total = ( 3 8 log log log 2 8 )=1.56. Fgure 1 llustrates a smple mx networ consstng of 3 mxes, 3 senders and 3 recevers. Assume the attacer does not have any a-pror nowledge about the senders and recevers except the traffc pattern. When all message flows are Posson processes wth the same rate, a message flow havng arrved at a mx was equally lely to have been forwarded to all of the possble next hops. Then the probablty dstrbuton of the output ln of a mx forwardng nput data flows s showed n the fgure. For example, the probablty dstrbuton of output 1 lns of mx 2 s {A : 4,B : 1 4,C : 1 2 }. We rewrte the probablty dstrbuton as { 1 4 A B + 1 2C}. Thus we can compute the prvacy entropy comng out of mx node 3 s: H3 total = ( 3 8 log log log 2 8 )=1.56, whch means the attacer needs 1.56 bts nformaton to dentfy who s the sender. Note that both mx 1 and mx 2 have contrbuted to the prvacy entropy besdes mx 3, and the fnal prvacy entropy s the synergstc effect of all three mxes Fg. 2. The Extended Networ for Prvacy Measurement. For each node M, an auxlary node M s created so as to support the cases where the mx nodes can be senders or recevers. It also contans a loop from M1 to M3. We extend the prvacy metrc to the networ model where the mxes can be senders or recevers and they can form message flow loops. Fg. 2 extends the smple mx networ n Fg. 1 n two aspects. Frst, we ntroduce an auxlary node nto the graph for each mx, and adds the same number of ncomng and outgong edges between the auxlary node and the mx node. Then the resultng graph s able to correctly descrbe cases n whch the mx nodes are senders or recevers. For nstance, the new graph s able to descrbe two sessons between B to M1 and M1 to M3 by B M1 M1 and M1 M1 M3 M3. Next, f there s a message flow loop from M3 to M1 n the graph as showed n Fg. 2, then we need to determne the probablty of the message flow over the loop. In ths case, we assume the message flow from M3 to M1 s X, then we can nfer the message flow over every ln as shown n Fg. 2. Snce the message flow from M3 to M1 s X, we can have 5 64 X B C M M M3=X, yeldng X = 1 59 (5B+4C +10M1+8M2+32M3). Then the prvacy entropy of each flow can be obtaned by the formula above. Then we can use ths prvacy measurement approach to evaluate the proposed scheme Prv-Code. Suppose the wreless networ conssts of N nodes and each node s transmttng data as an ndependent Posson process of the same transmsson rate. Then the prvacy entropy of an output data flow of a node can be computed from the probablty dstrbuton of the nput flows. V. PERFORMANCE ANALYSIS AND SIMULATION Computaton Overhead: The computaton overhead comes from two sources, the encodng procedure and the anonymous routng protocol. Computaton cost of the anonymous routng protocol s relatvely lghtweght, snce there s very few publc ey operatons n the protocol. Hence we manly focus on the encodng computaton overhead. The proposed scheme does not rely on expensve publc ey cryptographc mechansms to protect encodng vectors, ether. At each source/ntermedate node, a random codng
8 8 8 7 Sngle-Path MxNet Mult-Path MxNet Prv-Code Max Rate of Mult-Path MxNet Mult-Path MxNet Prv-Code Average Transmsson Rate Max Rate of Prv-Code Average Prvacy Entropy Sngle-Path MxNet Prv-Code and Mult-Path MxNet Average Transmsson Rate No. of Concurrent Sessons No. of Concurrent Sessons Ln Recepton Probablty (a) Average Transmsson Rate of Nodes vs. Sesson Number (b) Average Prvacy Entropy vs. Sesson Number (c) Average Prvacy Entropy vs. Ln Recepton Probablty Fg. 3. Prvacy Entropy and Transmsson Rate Comparson of Prv-Code, Sngle-Path and Mult-Path Mx Networ. The Networ sze s 50, each node has 5 neghbors on average, ln capacty s 100 unts, and each communcaton sesson s 2 unts. vector s generated and used to encode pacets from the same sesson. Then the codng vector of the newly generated pacet s attached to the pacet header. Ths procedure s much more effcent than encodng/decodng usng homomorphc encrypton. For each pacet receved at an ntermedate node, t needs to verfy whether t s ndependent from cached pacets. Gauss- Jordan elmnaton can be used to chec whether a pacet s lnearly ndependent, and the computaton complexty s O(n 3 ), where n s the sze of a generaton. Communcaton Overhead: For networ codng, the encodng coeffcents need to be put n the pacet header for the ntermedate nodes or the destnaton node to re-encode or decode pacets. Ths part of overhead results n addtonal communcaton cost. If the coeffcents obtan ther value from 0 to 255,.e., the sze of a byte, then the sze of coeffcents n a pacet header s n bytes. If the generaton sze s 20, that s, 20 pacets are grouped nto a generaton, then the networ codng coeffcent overhead s 20 bytes. For a pacet wth sze of 1000 bytes, the overhead s only 2% of the whole pacet sze. Storage Requrement: In the proposed scheme, the source node has to cache all pacets n the generaton before an acnowledgement s receved; the ntermedate nodes need to cache all receved lnearly ndependent pacets n the generaton before recevng an acnowledgement; the destnaton also has to cache all lnearly ndependent pacets n the same generaton before they can be decoded. Thus, the networ codng mechansm demands much more storage space than tradtonal transmsson technque. If the sze of a generaton s n, and on average there are m concurrent sessons passng through a node, then a node has to allocate O(nm) pacet cache on average. Note that the ntermedate nodes may not need nm pacet cache snce the pacets n a generaton are transmtted va multple paths. Ths requrement on storage can be tuned by settng the generaton sze n, and s also determned by the networ traffc. Smulaton and Evaluaton: We mplement our proposed networ codng schedulng scheme and conduct experments wth MatLab and NS2 to evaluate ts prvacy protecton capablty and performance. We use the prvacy metrc presented n Secton IV to compare our scheme Prv-Code wth two typcal mx networs, a sngle-path mx networ and a mult-path mx networ. The sngle-path mx networ always chooses the shortest path from a source to ts destnaton, whle the mult-path mx networ selects multple paths to transmt data and t uses a smlar optmzaton approach as Prv-Code. The mult-path mx networ s dfferent from Prv-Code n that transmssons over paths are ndependent of each other. Both mx networs for comparson are composed of exponental mxes that delay pacets accordng to the exponental dstrbuton, so the output flows are ndependent Posson processes. The wreless networ n our experments conssts 50 randomly deployed nodes wth node densty 6,.e., each node has 5 neghbors on average [7]. We assume that all lns have the same capacty, and nodes wthn the nterference range share the capacty. The ln capacty must guarantee the optmzaton algorthm has a optmal soluton. In our experments, we fx the ln capacty to be 100 unts. In our experments, we change the followng parameters n our experments: Concurrent sesson number: The number of concurrent sessons ranges from 4 to 20, whch means there are at most 20 pars of nodes are communcatng at the same tme. The communcaton rate of each sesson s 2 unts. Ln recepton probablty: The recepton probablty of each ln p Jj s a tunable parameter, whose dstrbuton conforms to the unform dstrbuton wth a mean rangng from 0.5 to 1.0. We compute the average prvacy entropy of all message flows based on the prvacy measurement method n Secton IV, and at the same tme, we compute the average transmsson rate and the maxmum transmsson rate among all nodes n the networ. Fg. 3(a) shows the average transmsson rate of the sngle-path mx networ, the mult-path mx networ and our scheme Prv-Code. Snce we have to nject dummy traffc nto the networ to mae each node have dentcal transmsson rate,.e. the maxmum transmsson rate, we also show the
9 9 maxmum transmsson rate for each experment n ths fgure. It can be seen that the requred transmsson rate of Prv-Code s between the sngle-path mx networ and the mult-path mx networ, whch accords wth our expectaton as Prv-Code tres to provde both strong prvacy and good performance. It means that Prv-Code can transmt the gven traffc wth less transmsson rate than the mult-path mx networ. Note that dummy traffc of Prv-Code to be njected nto the networ s also less than that of the mult-path mx networ. We show n Fg. 3(b) the average prvacy entropy over all nodes provded by Prv-Code, the sngle-path mx networ and the mult-path mx networ. Snce we mae each node have dentcal transmsson rate by traffc paddng, Prv-Code and the mult-path mx networ have the same prvacy entropy. In contrast, the sngle-path mx networ always chooses the shortest path for data transmsson, the prvacy changes wth the number of concurrent sessons n the networ. When there are fewer sessons, t s harder for the sngle-path mx networ to protect ther prvacy, but the prvacy entropy grows as the number of sessons ncreases. Fg. 3(c) shows the average transmsson rate versus dfferent ln recepton probablty for Prv-Code and the mult-path mx networ. It shows that the requred transmsson rate of both Prv-Code and the mult-path mx networ decreases as the ln qualty becomes better. But the transmsson rate of Prv-Code s almost half of that of the mult-path mx networ, whch means Prv-Code can transmt the same amount of traffc wth about half of the transmsson rate compared wth the mult-path mx networ. demonstrates the great advantage of networ codng n performance mprovement. VI. CONCLUSION In ths paper, we nvestgate the problem of how to explot networ codng to protect prvacy aganst traffc analyss attacs under a powerful threat model, n whch the attacer s able to contnuously montor the entre networ and mount both passve and actve attacs aganst the wreless networ. Based on nformaton theory, we formally defne the prvacy entropy n terms of traffc flow nformaton. Then we formalze a hypergraph-based networ model for wreless networs based on networ codng, and formulate an optmzaton problem to see the optmal networ codng schedulng. After that, we provde a decentralzed algorthm to solve the optmzaton problem. Our analyss and experment evaluaton show that the proposed scheme has substantal advantage over exstng schemes on prvacy protecton aganst traffc analyss. ACKNOWLEDGMENTS Zhguo s study s supported n part by Scentfc Foundaton for Returned Overseas Chnese Scholars, MOE, and the NSFC project under Grant No Ka s research s supported by NSFC under grant , Jangsu Natural Scence Foundaton under grant BK , and Natonal Basc Research Program of Chna (973) under grant 2011CB Yunhao s study s supported n part by NSFC Dstngushed Young Scholars Program under Grant , and Natonal Hgh-Tech R&D Program of Chna (863) under grant No. 2011AA REFERENCES [1] D. Chaum, Untraceable electronc mal, return addresses, and dgtal pseudonyms, Communcatons of the ACM, vol. 4, no. 2, February [2] M. K. Reter and A. D. Rubn, Crowds: Anonymty for web transactons, ACM Trans. Inf. Syst. Secur., vol. 1, no. 1, pp , [3] R. Dngledne, N. Mathewson, and P. F. Syverson, Tor: The secondgeneraton onon router, n USENIX Securty Symposum, 2004, pp [4] U. Moller and L. Cottrell, Mxmaster protocol verson 2, IETF Internet draft, [5] G. Danezs, R. Dngledne, and N. Mathewson, Mxmnon: Desgn of a type anonymous remaler protocol, n IEEE Symposum on Securty and Prvacy, 2003, pp [6] R. Ahlswede, N. Ca, S.-Y. R. L, and R. W. Yeung, Networ nformaton flow, IEEE Trans. Inf. Theory, vol. 46, no. 4, pp , [7] Y. Fan, Y. Jang, H. Zhu, and X. Shen, An effcent prvacy-preservng scheme aganst traffc analyss attacs n networ codng, n INFO- COM, 2009, pp [8] T. Cu, L. Chen, and T. Ho, Energy effcent opportunstc networ codng for wreless networs, n INFOCOM, 2008, pp [9] G. Danezs and I. Goldberg, Sphnx: A compact and provably secure mx format, n IEEE Symposum on Securty and Prvacy, 2009, pp [10] A. Serjantov and G. Danezs, Towards an nformaton theoretc metrc for anonymty, n Prvacy Enhancng Technologes, 2002, pp [11] S. Katt, H. Rahul, W. Hu, D. Katab, M. Médard, and J. Crowcroft, Xors n the ar: practcal wreless networ codng, n SIGCOMM, 2006, pp [12] M. N. Krohn, M. J. Freedman, and D. Mazères, On-the-fly verfcaton of rateless erasure codes for effcent content dstrbuton, n IEEE Symposum on Securty and Prvacy, 2004, pp [13] Z. Yu, Y. We, B. Ramumar, and Y. Guan, An effcent scheme for securng xor networ codng aganst polluton attacs, n INFOCOM, 2009, pp [14] X. Chang, J. Wang, J. Wang, V. Lee, K. Lu, and Y. Yang, On achevng maxmum secure throughput usng networ codng aganst wretap attac, n ICDCS, 2010, pp [15] J. Wang, J. Wang, K. Lu, B. Xao, and N. Gu, Optmal lnear networ codng desgn for secure uncast wth multple streams, n INFOCOM, 2010, pp [16] S. Jagg, M. Langberg, S. Katt, T. Ho, D. Katab, M. Médard, and M. Effros, Reslent networ codng n the presence of byzantne adversares, IEEE Transactons on Informaton Theory, vol. 54, no. 6, pp , [17] P. Zhang, Y. Jang, C. Ln, Y. Fan, and X. Shen, P-codng: Secure networ codng aganst eavesdroppng attacs, n INFOCOM, 2010, pp [18] Y. Jang, Y. Fan, X. S. Shen, and C. Ln, A self-adaptve probablstc pacet flterng scheme aganst entropy attacs n networ codng, Computer Networs, vol. 53, no. 18, pp , [19] Z. Wan, K. Ren, B. Zhu, B. Preneel, and M. Gu, Anonymous user communcaton for prvacy protecton n wreless metropoltan mesh networs, IEEE Trans. Veh. Technol., no. 2, pp , [20] G. Danezs, The traffc analyss of contnuous-tme mxes, n PET04, [21] Y. Zhu, X. Fu, B. Graham, R. Bettat, and W. Zhao, On flow correlaton attacs and countermeasures n mx networs, n PET04, LNCS 3424, 2004, pp [22] X. Zhang and B. L, Optmzed multpath networ codng n lossy wreless networs, IEEE Journal on Selected Areas n Communcatons, vol. 27, no. 5, pp , [23] D. P. Palomar and M. Chang, A tutoral on decomposton methods for networ utlty maxmzaton, IEEE Journal on Selected Areas n Communcatons, vol. 24, no. 8, pp , [24] D. P. Bertseas and J. Tstsls, Parallel and Dstrbuted Computaton: Numercal Methods. Prentce-Hall, Inc., [25] V. Shmatov and M.-H. Wang, Tmng analyss n low-latency mx networs: Attacs and defenses, n ESORICS, 2006, pp
denote the location of a node, and suppose node X . This transmission causes a successful reception by node X for any other node
Fnal Report of EE359 Class Proect Throughput and Delay n Wreless Ad Hoc Networs Changhua He changhua@stanford.edu Abstract: Networ throughput and pacet delay are the two most mportant parameters to evaluate
More informationMinimal Coding Network With Combinatorial Structure For Instantaneous Recovery From Edge Failures
Mnmal Codng Network Wth Combnatoral Structure For Instantaneous Recovery From Edge Falures Ashly Joseph 1, Mr.M.Sadsh Sendl 2, Dr.S.Karthk 3 1 Fnal Year ME CSE Student Department of Computer Scence Engneerng
More informationThe Development of Web Log Mining Based on Improve-K-Means Clustering Analysis
The Development of Web Log Mnng Based on Improve-K-Means Clusterng Analyss TngZhong Wang * College of Informaton Technology, Luoyang Normal Unversty, Luoyang, 471022, Chna wangtngzhong2@sna.cn Abstract.
More informationRelay Secrecy in Wireless Networks with Eavesdropper
Relay Secrecy n Wreless Networks wth Eavesdropper Parvathnathan Venktasubramanam, Tng He and Lang Tong School of Electrcal and Computer Engneerng Cornell Unversty, Ithaca, NY 14853 Emal : {pv45, th255,
More informationA Secure Password-Authenticated Key Agreement Using Smart Cards
A Secure Password-Authentcated Key Agreement Usng Smart Cards Ka Chan 1, Wen-Chung Kuo 2 and Jn-Chou Cheng 3 1 Department of Computer and Informaton Scence, R.O.C. Mltary Academy, Kaohsung 83059, Tawan,
More informationModule 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur
Module LOSSLESS IMAGE COMPRESSION SYSTEMS Lesson 3 Lossless Compresson: Huffman Codng Instructonal Objectves At the end of ths lesson, the students should be able to:. Defne and measure source entropy..
More informationPerformance Analysis of Energy Consumption of Smartphone Running Mobile Hotspot Application
Internatonal Journal of mart Grd and lean Energy Performance Analyss of Energy onsumpton of martphone Runnng Moble Hotspot Applcaton Yun on hung a chool of Electronc Engneerng, oongsl Unversty, 511 angdo-dong,
More informationAn Alternative Way to Measure Private Equity Performance
An Alternatve Way to Measure Prvate Equty Performance Peter Todd Parlux Investment Technology LLC Summary Internal Rate of Return (IRR) s probably the most common way to measure the performance of prvate
More informationbenefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).
REVIEW OF RISK MANAGEMENT CONCEPTS LOSS DISTRIBUTIONS AND INSURANCE Loss and nsurance: When someone s subject to the rsk of ncurrng a fnancal loss, the loss s generally modeled usng a random varable or
More informationAPPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT
APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT Toshhko Oda (1), Kochro Iwaoka (2) (1), (2) Infrastructure Systems Busness Unt, Panasonc System Networks Co., Ltd. Saedo-cho
More informationProactive Secret Sharing Or: How to Cope With Perpetual Leakage
Proactve Secret Sharng Or: How to Cope Wth Perpetual Leakage Paper by Amr Herzberg Stanslaw Jareck Hugo Krawczyk Mot Yung Presentaton by Davd Zage What s Secret Sharng Basc Idea ((2, 2)-threshold scheme):
More informationAn Adaptive and Distributed Clustering Scheme for Wireless Sensor Networks
2007 Internatonal Conference on Convergence Informaton Technology An Adaptve and Dstrbuted Clusterng Scheme for Wreless Sensor Networs Xnguo Wang, Xnmng Zhang, Guolang Chen, Shuang Tan Department of Computer
More informationHow To Improve Delay Throughput In Wireless Networks With Multipath Routing And Channel Codeing
Delay-Throughput Enhancement n Wreless Networs wth Mult-path Routng and Channel Codng Kevan Ronas, Student Member, IEEE, Amr-Hamed Mohsenan-Rad, Member, IEEE, Vncent W.S. Wong, Senor Member, IEEE, Sathsh
More informationData Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network *
JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 24, 819-840 (2008) Data Broadcast on a Mult-System Heterogeneous Overlayed Wreless Network * Department of Computer Scence Natonal Chao Tung Unversty Hsnchu,
More informationSupport Vector Machines
Support Vector Machnes Max Wellng Department of Computer Scence Unversty of Toronto 10 Kng s College Road Toronto, M5S 3G5 Canada wellng@cs.toronto.edu Abstract Ths s a note to explan support vector machnes.
More informationA Design Method of High-availability and Low-optical-loss Optical Aggregation Network Architecture
A Desgn Method of Hgh-avalablty and Low-optcal-loss Optcal Aggregaton Network Archtecture Takehro Sato, Kuntaka Ashzawa, Kazumasa Tokuhash, Dasuke Ish, Satoru Okamoto and Naoak Yamanaka Dept. of Informaton
More informationTHE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek
HE DISRIBUION OF LOAN PORFOLIO VALUE * Oldrch Alfons Vascek he amount of captal necessary to support a portfolo of debt securtes depends on the probablty dstrbuton of the portfolo loss. Consder a portfolo
More informationLuby s Alg. for Maximal Independent Sets using Pairwise Independence
Lecture Notes for Randomzed Algorthms Luby s Alg. for Maxmal Independent Sets usng Parwse Independence Last Updated by Erc Vgoda on February, 006 8. Maxmal Independent Sets For a graph G = (V, E), an ndependent
More informationForecasting the Direction and Strength of Stock Market Movement
Forecastng the Drecton and Strength of Stock Market Movement Jngwe Chen Mng Chen Nan Ye cjngwe@stanford.edu mchen5@stanford.edu nanye@stanford.edu Abstract - Stock market s one of the most complcated systems
More information1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP)
6.3 / -- Communcaton Networks II (Görg) SS20 -- www.comnets.un-bremen.de Communcaton Networks II Contents. Fundamentals of probablty theory 2. Emergence of communcaton traffc 3. Stochastc & Markovan Processes
More informationRobust Design of Public Storage Warehouses. Yeming (Yale) Gong EMLYON Business School
Robust Desgn of Publc Storage Warehouses Yemng (Yale) Gong EMLYON Busness School Rene de Koster Rotterdam school of management, Erasmus Unversty Abstract We apply robust optmzaton and revenue management
More informationFeature selection for intrusion detection. Slobodan Petrović NISlab, Gjøvik University College
Feature selecton for ntruson detecton Slobodan Petrovć NISlab, Gjøvk Unversty College Contents The feature selecton problem Intruson detecton Traffc features relevant for IDS The CFS measure The mrmr measure
More informationFault tolerance in cloud technologies presented as a service
Internatonal Scentfc Conference Computer Scence 2015 Pavel Dzhunev, PhD student Fault tolerance n cloud technologes presented as a servce INTRODUCTION Improvements n technques for vrtualzaton and performance
More informationRecurrence. 1 Definitions and main statements
Recurrence 1 Defntons and man statements Let X n, n = 0, 1, 2,... be a MC wth the state space S = (1, 2,...), transton probabltes p j = P {X n+1 = j X n = }, and the transton matrx P = (p j ),j S def.
More informationAn Interest-Oriented Network Evolution Mechanism for Online Communities
An Interest-Orented Network Evoluton Mechansm for Onlne Communtes Cahong Sun and Xaopng Yang School of Informaton, Renmn Unversty of Chna, Bejng 100872, P.R. Chna {chsun,yang}@ruc.edu.cn Abstract. Onlne
More informationHow Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence
1 st Internatonal Symposum on Imprecse Probabltes and Ther Applcatons, Ghent, Belgum, 29 June 2 July 1999 How Sets of Coherent Probabltes May Serve as Models for Degrees of Incoherence Mar J. Schervsh
More informationAnalysis of Energy-Conserving Access Protocols for Wireless Identification Networks
From the Proceedngs of Internatonal Conference on Telecommuncaton Systems (ITC-97), March 2-23, 1997. 1 Analyss of Energy-Conservng Access Protocols for Wreless Identfcaton etworks Imrch Chlamtac a, Chara
More informationProject Networks With Mixed-Time Constraints
Project Networs Wth Mxed-Tme Constrants L Caccetta and B Wattananon Western Australan Centre of Excellence n Industral Optmsaton (WACEIO) Curtn Unversty of Technology GPO Box U1987 Perth Western Australa
More informationJ. Parallel Distrib. Comput.
J. Parallel Dstrb. Comput. 71 (2011) 62 76 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. journal homepage: www.elsever.com/locate/jpdc Optmzng server placement n dstrbuted systems n
More informationPAS: A Packet Accounting System to Limit the Effects of DoS & DDoS. Debish Fesehaye & Klara Naherstedt University of Illinois-Urbana Champaign
PAS: A Packet Accountng System to Lmt the Effects of DoS & DDoS Debsh Fesehaye & Klara Naherstedt Unversty of Illnos-Urbana Champagn DoS and DDoS DDoS attacks are ncreasng threats to our dgtal world. Exstng
More informationINVESTIGATION OF VEHICULAR USERS FAIRNESS IN CDMA-HDR NETWORKS
21 22 September 2007, BULGARIA 119 Proceedngs of the Internatonal Conference on Informaton Technologes (InfoTech-2007) 21 st 22 nd September 2007, Bulgara vol. 2 INVESTIGATION OF VEHICULAR USERS FAIRNESS
More informationOn File Delay Minimization for Content Uploading to Media Cloud via Collaborative Wireless Network
On Fle Delay Mnmzaton for Content Uploadng to Meda Cloud va Collaboratve Wreless Network Ge Zhang and Yonggang Wen School of Computer Engneerng Nanyang Technologcal Unversty Sngapore Emal: {zh0001ge, ygwen}@ntu.edu.sg
More informationAN APPOINTMENT ORDER OUTPATIENT SCHEDULING SYSTEM THAT IMPROVES OUTPATIENT EXPERIENCE
AN APPOINTMENT ORDER OUTPATIENT SCHEDULING SYSTEM THAT IMPROVES OUTPATIENT EXPERIENCE Yu-L Huang Industral Engneerng Department New Mexco State Unversty Las Cruces, New Mexco 88003, U.S.A. Abstract Patent
More informationForecasting the Demand of Emergency Supplies: Based on the CBR Theory and BP Neural Network
700 Proceedngs of the 8th Internatonal Conference on Innovaton & Management Forecastng the Demand of Emergency Supples: Based on the CBR Theory and BP Neural Network Fu Deqang, Lu Yun, L Changbng School
More informationA hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm
Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(7):1884-1889 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A hybrd global optmzaton algorthm based on parallel
More informationAd-Hoc Games and Packet Forwardng Networks
On Desgnng Incentve-Compatble Routng and Forwardng Protocols n Wreless Ad-Hoc Networks An Integrated Approach Usng Game Theoretcal and Cryptographc Technques Sheng Zhong L (Erran) L Yanbn Grace Lu Yang
More informationPower-of-Two Policies for Single- Warehouse Multi-Retailer Inventory Systems with Order Frequency Discounts
Power-of-wo Polces for Sngle- Warehouse Mult-Retaler Inventory Systems wth Order Frequency Dscounts José A. Ventura Pennsylvana State Unversty (USA) Yale. Herer echnon Israel Insttute of echnology (Israel)
More informationOn the Interaction between Load Balancing and Speed Scaling
On the Interacton between Load Balancng and Speed Scalng Ljun Chen and Na L Abstract Speed scalng has been wdely adopted n computer and communcaton systems, n partcular, to reduce energy consumpton. An
More informationA New Task Scheduling Algorithm Based on Improved Genetic Algorithm
A New Task Schedulng Algorthm Based on Improved Genetc Algorthm n Cloud Computng Envronment Congcong Xong, Long Feng, Lxan Chen A New Task Schedulng Algorthm Based on Improved Genetc Algorthm n Cloud Computng
More informationA 2 -MAC: An Adaptive, Anycast MAC Protocol for Wireless Sensor Networks
: An Adaptve, Anycast MAC Protocol for Wreless Sensor Networks Hwee-Xan Tan and Mun Choon Chan Department of Computer Scence, School of Computng, Natonal Unversty of Sngapore {hweexan, chanmc}@comp.nus.edu.sg
More informationANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING
ANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING Matthew J. Lberatore, Department of Management and Operatons, Vllanova Unversty, Vllanova, PA 19085, 610-519-4390,
More informationEfficient Bandwidth Management in Broadband Wireless Access Systems Using CAC-based Dynamic Pricing
Effcent Bandwdth Management n Broadband Wreless Access Systems Usng CAC-based Dynamc Prcng Bader Al-Manthar, Ndal Nasser 2, Najah Abu Al 3, Hossam Hassanen Telecommuncatons Research Laboratory School of
More information2008/8. An integrated model for warehouse and inventory planning. Géraldine Strack and Yves Pochet
2008/8 An ntegrated model for warehouse and nventory plannng Géraldne Strack and Yves Pochet CORE Voe du Roman Pays 34 B-1348 Louvan-la-Neuve, Belgum. Tel (32 10) 47 43 04 Fax (32 10) 47 43 01 E-mal: corestat-lbrary@uclouvan.be
More informationWatermark-based Provable Data Possession for Multimedia File in Cloud Storage
Vol.48 (CIA 014), pp.103-107 http://dx.do.org/10.1457/astl.014.48.18 Watermar-based Provable Data Possesson for Multmeda Fle n Cloud Storage Yongjun Ren 1,, Jang Xu 1,, Jn Wang 1,, Lmng Fang 3, Jeong-U
More informationWhen Network Effect Meets Congestion Effect: Leveraging Social Services for Wireless Services
When Network Effect Meets Congeston Effect: Leveragng Socal Servces for Wreless Servces aowen Gong School of Electrcal, Computer and Energy Engeerng Arzona State Unversty Tempe, AZ 8587, USA xgong9@asuedu
More informationEnergy Conserving Routing in Wireless Ad-hoc Networks
Energy Conservng Routng n Wreless Ad-hoc Networks Jae-Hwan Chang and Leandros Tassulas Department of Electrcal and Computer Engneerng & Insttute for Systems Research Unversty of Maryland at College ark
More informationEnabling P2P One-view Multi-party Video Conferencing
Enablng P2P One-vew Mult-party Vdeo Conferencng Yongxang Zhao, Yong Lu, Changja Chen, and JanYn Zhang Abstract Mult-Party Vdeo Conferencng (MPVC) facltates realtme group nteracton between users. Whle P2P
More informationThe Greedy Method. Introduction. 0/1 Knapsack Problem
The Greedy Method Introducton We have completed data structures. We now are gong to look at algorthm desgn methods. Often we are lookng at optmzaton problems whose performance s exponental. For an optmzaton
More informationEfficient On-Demand Data Service Delivery to High-Speed Trains in Cellular/Infostation Integrated Networks
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. XX, NO. XX, MONTH 2XX 1 Effcent On-Demand Data Servce Delvery to Hgh-Speed Trans n Cellular/Infostaton Integrated Networks Hao Lang, Student Member,
More informationSecure Network Coding Over the Integers
Secure Network Codng Over the Integers Rosaro Gennaro Jonathan Katz Hugo Krawczyk Tal Rabn Abstract Network codng has receved sgnfcant attenton n the networkng communty for ts potental to ncrease throughput
More informationOn the Optimal Control of a Cascade of Hydro-Electric Power Stations
On the Optmal Control of a Cascade of Hydro-Electrc Power Statons M.C.M. Guedes a, A.F. Rbero a, G.V. Smrnov b and S. Vlela c a Department of Mathematcs, School of Scences, Unversty of Porto, Portugal;
More informationDynamic Pricing for Smart Grid with Reinforcement Learning
Dynamc Prcng for Smart Grd wth Renforcement Learnng Byung-Gook Km, Yu Zhang, Mhaela van der Schaar, and Jang-Won Lee Samsung Electroncs, Suwon, Korea Department of Electrcal Engneerng, UCLA, Los Angeles,
More informationMulti-Source Video Multicast in Peer-to-Peer Networks
ult-source Vdeo ultcast n Peer-to-Peer Networks Francsco de Asís López-Fuentes*, Eckehard Stenbach Technsche Unverstät ünchen Insttute of Communcaton Networks, eda Technology Group 80333 ünchen, Germany
More informationDistributed Optimal Contention Window Control for Elastic Traffic in Wireless LANs
Dstrbuted Optmal Contenton Wndow Control for Elastc Traffc n Wreless LANs Yalng Yang, Jun Wang and Robn Kravets Unversty of Illnos at Urbana-Champagn { yyang8, junwang3, rhk@cs.uuc.edu} Abstract Ths paper
More information8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by
6 CHAPTER 8 COMPLEX VECTOR SPACES 5. Fnd the kernel of the lnear transformaton gven n Exercse 5. In Exercses 55 and 56, fnd the mage of v, for the ndcated composton, where and are gven by the followng
More informationAn MILP model for planning of batch plants operating in a campaign-mode
An MILP model for plannng of batch plants operatng n a campagn-mode Yanna Fumero Insttuto de Desarrollo y Dseño CONICET UTN yfumero@santafe-concet.gov.ar Gabrela Corsano Insttuto de Desarrollo y Dseño
More informationHow To Detect An 802.11 Traffc From A Network With A Network Onlne Onlnet
IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. X, NO. X, XXX 2008 1 Passve Onlne Detecton of 802.11 Traffc Usng Sequental Hypothess Testng wth TCP ACK-Pars We We, Member, IEEE, Kyoungwon Suh, Member, IEEE,
More informationPerformance Analysis and Comparison of QoS Provisioning Mechanisms for CBR Traffic in Noisy IEEE 802.11e WLANs Environments
Tamkang Journal of Scence and Engneerng, Vol. 12, No. 2, pp. 143149 (2008) 143 Performance Analyss and Comparson of QoS Provsonng Mechansms for CBR Traffc n Nosy IEEE 802.11e WLANs Envronments Der-Junn
More informationWireless Inter-Session Network Coding - An Approach Using Virtual Multicasts
Wreless Inter-Sesson Network Codng - An Approach Usng Vrtual Multcasts Mchael Hendlmaer,DesmondS.Lun,DanalTraskov,andMurelMédard Insttute for Communcatons Engneerng CCIB RLE Technsche Unverstät München
More informationStudy on Model of Risks Assessment of Standard Operation in Rural Power Network
Study on Model of Rsks Assessment of Standard Operaton n Rural Power Network Qngj L 1, Tao Yang 2 1 Qngj L, College of Informaton and Electrcal Engneerng, Shenyang Agrculture Unversty, Shenyang 110866,
More informationOn the Interaction between Load Balancing and Speed Scaling
On the Interacton between Load Balancng and Speed Scalng Ljun Chen, Na L and Steven H. Low Engneerng & Appled Scence Dvson, Calforna Insttute of Technology, USA Abstract Speed scalng has been wdely adopted
More informationInstitute of Informatics, Faculty of Business and Management, Brno University of Technology,Czech Republic
Lagrange Multplers as Quanttatve Indcators n Economcs Ivan Mezník Insttute of Informatcs, Faculty of Busness and Management, Brno Unversty of TechnologCzech Republc Abstract The quanttatve role of Lagrange
More informationAn Ad Hoc Network Load Balancing Energy- Efficient Multipath Routing Protocol
246 JOURNA OF SOFTWAR, VO. 9, NO. 1, JANUARY 2014 An Ad Hoc Network oad alancng nergy- ffcent Multpath Routng Protocol De-jn Kong Shanx Fnance and Taxaton College, Tayuan, Chna mal: dejnkong@163.com Xao-lng
More informationA Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression
Novel Methodology of Workng Captal Management for Large Publc Constructons by Usng Fuzzy S-curve Regresson Cheng-Wu Chen, Morrs H. L. Wang and Tng-Ya Hseh Department of Cvl Engneerng, Natonal Central Unversty,
More informationFrequency Selective IQ Phase and IQ Amplitude Imbalance Adjustments for OFDM Direct Conversion Transmitters
Frequency Selectve IQ Phase and IQ Ampltude Imbalance Adjustments for OFDM Drect Converson ransmtters Edmund Coersmeer, Ernst Zelnsk Noka, Meesmannstrasse 103, 44807 Bochum, Germany edmund.coersmeer@noka.com,
More informationRisk-based Fatigue Estimate of Deep Water Risers -- Course Project for EM388F: Fracture Mechanics, Spring 2008
Rsk-based Fatgue Estmate of Deep Water Rsers -- Course Project for EM388F: Fracture Mechancs, Sprng 2008 Chen Sh Department of Cvl, Archtectural, and Envronmental Engneerng The Unversty of Texas at Austn
More informationtaposh_kuet20@yahoo.comcsedchan@cityu.edu.hk rajib_csedept@yahoo.co.uk, alam_shihabul@yahoo.com
G. G. Md. Nawaz Al 1,2, Rajb Chakraborty 2, Md. Shhabul Alam 2 and Edward Chan 1 1 Cty Unversty of Hong Kong, Hong Kong, Chna taposh_kuet20@yahoo.comcsedchan@ctyu.edu.hk 2 Khulna Unversty of Engneerng
More informationOpen Access A Load Balancing Strategy with Bandwidth Constraint in Cloud Computing. Jing Deng 1,*, Ping Guo 2, Qi Li 3, Haizhu Chen 1
Send Orders for Reprnts to reprnts@benthamscence.ae The Open Cybernetcs & Systemcs Journal, 2014, 8, 115-121 115 Open Access A Load Balancng Strategy wth Bandwdth Constrant n Cloud Computng Jng Deng 1,*,
More informationVoIP Playout Buffer Adjustment using Adaptive Estimation of Network Delays
VoIP Playout Buffer Adjustment usng Adaptve Estmaton of Network Delays Mroslaw Narbutt and Lam Murphy* Department of Computer Scence Unversty College Dubln, Belfeld, Dubln, IRELAND Abstract The poor qualty
More informationEnergy Efficient Routing in Ad Hoc Disaster Recovery Networks
Energy Effcent Routng n Ad Hoc Dsaster Recovery Networks Gl Zussman and Adran Segall Department of Electrcal Engneerng Technon Israel Insttute of Technology Hafa 32000, Israel {glz@tx, segall@ee}.technon.ac.l
More informationVRT012 User s guide V0.1. Address: Žirmūnų g. 27, Vilnius LT-09105, Phone: (370-5) 2127472, Fax: (370-5) 276 1380, Email: info@teltonika.
VRT012 User s gude V0.1 Thank you for purchasng our product. We hope ths user-frendly devce wll be helpful n realsng your deas and brngng comfort to your lfe. Please take few mnutes to read ths manual
More informationEconomic-Robust Transmission Opportunity Auction in Multi-hop Wireless Networks
Economc-Robust Transmsson Opportunty Aucton n Mult-hop Wreless Networks Mng L, Pan L, Mao Pan, and Jnyuan Sun Department of Electrcal and Computer Engneerng, Msssspp State Unversty, Msssspp State, MS 39762
More information"Research Note" APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES *
Iranan Journal of Scence & Technology, Transacton B, Engneerng, ol. 30, No. B6, 789-794 rnted n The Islamc Republc of Iran, 006 Shraz Unversty "Research Note" ALICATION OF CHARGE SIMULATION METHOD TO ELECTRIC
More informationA Lyapunov Optimization Approach to Repeated Stochastic Games
PROC. ALLERTON CONFERENCE ON COMMUNICATION, CONTROL, AND COMPUTING, OCT. 2013 1 A Lyapunov Optmzaton Approach to Repeated Stochastc Games Mchael J. Neely Unversty of Southern Calforna http://www-bcf.usc.edu/
More informationLAMOR: Lifetime-Aware Multipath Optimized Routing Algorithm for Video Transmission over Ad Hoc Networks
LAMOR: Lfetme-Aware Multpath Optmzed Routng Algorthm for Vdeo ransmsson over Ad Hoc Networks 1 Lansheng an, Lng Xe, Kng-m Ko, Mng Le and Moshe Zukerman Abstract Multpath routng s a key technque to support
More informationDEFINING %COMPLETE IN MICROSOFT PROJECT
CelersSystems DEFINING %COMPLETE IN MICROSOFT PROJECT PREPARED BY James E Aksel, PMP, PMI-SP, MVP For Addtonal Informaton about Earned Value Management Systems and reportng, please contact: CelersSystems,
More informationApplication of Multi-Agents for Fault Detection and Reconfiguration of Power Distribution Systems
1 Applcaton of Mult-Agents for Fault Detecton and Reconfguraton of Power Dstrbuton Systems K. Nareshkumar, Member, IEEE, M. A. Choudhry, Senor Member, IEEE, J. La, A. Felach, Senor Member, IEEE Abstract--The
More informationDownlink Power Allocation for Multi-class. Wireless Systems
Downlnk Power Allocaton for Mult-class 1 Wreless Systems Jang-Won Lee, Rav R. Mazumdar, and Ness B. Shroff School of Electrcal and Computer Engneerng Purdue Unversty West Lafayette, IN 47907, USA {lee46,
More informationIn some supply chains, materials are ordered periodically according to local information. This paper investigates
MANUFACTURING & SRVIC OPRATIONS MANAGMNT Vol. 12, No. 3, Summer 2010, pp. 430 448 ssn 1523-4614 essn 1526-5498 10 1203 0430 nforms do 10.1287/msom.1090.0277 2010 INFORMS Improvng Supply Chan Performance:
More informationDynamic Fleet Management for Cybercars
Proceedngs of the IEEE ITSC 2006 2006 IEEE Intellgent Transportaton Systems Conference Toronto, Canada, September 17-20, 2006 TC7.5 Dynamc Fleet Management for Cybercars Fenghu. Wang, Mng. Yang, Ruqng.
More informationMethodology to Determine Relationships between Performance Factors in Hadoop Cloud Computing Applications
Methodology to Determne Relatonshps between Performance Factors n Hadoop Cloud Computng Applcatons Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng and
More informationWhat is Candidate Sampling
What s Canddate Samplng Say we have a multclass or mult label problem where each tranng example ( x, T ) conssts of a context x a small (mult)set of target classes T out of a large unverse L of possble
More informationMaster s Thesis. Configuring robust virtual wireless sensor networks for Internet of Things inspired by brain functional networks
Master s Thess Ttle Confgurng robust vrtual wreless sensor networks for Internet of Thngs nspred by bran functonal networks Supervsor Professor Masayuk Murata Author Shnya Toyonaga February 10th, 2014
More informationAn RFID Distance Bounding Protocol
An RFID Dstance Boundng Protocol Gerhard P. Hancke and Markus G. Kuhn May 22, 2006 An RFID Dstance Boundng Protocol p. 1 Dstance boundng Verfer d Prover Places an upper bound on physcal dstance Does not
More informationStochastic Games on a Multiple Access Channel
Stochastc Games on a Multple Access Channel Prashant N and Vnod Sharma Department of Electrcal Communcaton Engneerng Indan Insttute of Scence, Bangalore 560012, Inda Emal: prashant2406@gmal.com, vnod@ece.sc.ernet.n
More informationA Performance Analysis of View Maintenance Techniques for Data Warehouses
A Performance Analyss of Vew Mantenance Technques for Data Warehouses Xng Wang Dell Computer Corporaton Round Roc, Texas Le Gruenwald The nversty of Olahoma School of Computer Scence orman, OK 739 Guangtao
More informationAn Efficient Recovery Algorithm for Coverage Hole in WSNs
An Effcent Recover Algorthm for Coverage Hole n WSNs Song Ja 1,*, Wang Balng 1, Peng Xuan 1 School of Informaton an Electrcal Engneerng Harbn Insttute of Technolog at Weha, Shanong, Chna Automatc Test
More informationivoip: an Intelligent Bandwidth Management Scheme for VoIP in WLANs
VoIP: an Intellgent Bandwdth Management Scheme for VoIP n WLANs Zhenhu Yuan and Gabrel-Mro Muntean Abstract Voce over Internet Protocol (VoIP) has been wdely used by many moble consumer devces n IEEE 802.11
More informationReinforcement Learning for Quality of Service in Mobile Ad Hoc Network (MANET)
Renforcement Learnng for Qualty of Servce n Moble Ad Hoc Network (MANET) *T.KUMANAN AND **K.DURAISWAMY *Meenaksh College of Engneerng West K.K Nagar, Cheena-78 **Dean/academc,K.S.R College of Technology,Truchengode
More informationRequIn, a tool for fast web traffic inference
RequIn, a tool for fast web traffc nference Olver aul, Jean Etenne Kba GET/INT, LOR Department 9 rue Charles Fourer 90 Evry, France Olver.aul@nt-evry.fr, Jean-Etenne.Kba@nt-evry.fr Abstract As networked
More informationIMPACT ANALYSIS OF A CELLULAR PHONE
4 th ASA & μeta Internatonal Conference IMPACT AALYSIS OF A CELLULAR PHOE We Lu, 2 Hongy L Bejng FEAonlne Engneerng Co.,Ltd. Bejng, Chna ABSTRACT Drop test smulaton plays an mportant role n nvestgatng
More informationLogistic Regression. Lecture 4: More classifiers and classes. Logistic regression. Adaboost. Optimization. Multiple class classification
Lecture 4: More classfers and classes C4B Machne Learnng Hlary 20 A. Zsserman Logstc regresson Loss functons revsted Adaboost Loss functons revsted Optmzaton Multple class classfcaton Logstc Regresson
More informationQoS-Aware Active Queue Management for Multimedia Services over the Internet
QoS-Aware Actve Queue Management for Multmeda Servces over the Internet I-Shyan Hwang, *Bor-Junn Hwang, Pen-Mng Chang, Cheng-Yu Wang Abstract Recently, the multmeda servces such as IPTV, vdeo conference
More informationAN APPROACH TO WIRELESS SCHEDULING CONSIDERING REVENUE AND USERS SATISFACTION
The Medterranean Journal of Computers and Networks, Vol. 2, No. 1, 2006 57 AN APPROACH TO WIRELESS SCHEDULING CONSIDERING REVENUE AND USERS SATISFACTION L. Bada 1,*, M. Zorz 2 1 Department of Engneerng,
More informationA New Paradigm for Load Balancing in Wireless Mesh Networks
A New Paradgm for Load Balancng n Wreless Mesh Networks Abstract: Obtanng maxmum throughput across a network or a mesh through optmal load balancng s known to be an NP-hard problem. Desgnng effcent load
More informationPeriod and Deadline Selection for Schedulability in Real-Time Systems
Perod and Deadlne Selecton for Schedulablty n Real-Tme Systems Thdapat Chantem, Xaofeng Wang, M.D. Lemmon, and X. Sharon Hu Department of Computer Scence and Engneerng, Department of Electrcal Engneerng
More informationCase Study: Load Balancing
Case Study: Load Balancng Thursday, 01 June 2006 Bertol Marco A.A. 2005/2006 Dmensonamento degl mpant Informatc LoadBal - 1 Introducton Optmze the utlzaton of resources to reduce the user response tme
More informationJoint Dynamic Radio Resource Allocation and Mobility Load Balancing in 3GPP LTE Multi-Cell Network
288 FENG LI, LINA GENG, SHIHUA ZHU, JOINT DYNAMIC RADIO RESOURCE ALLOCATION AND MOBILITY LOAD BALANCING Jont Dynamc Rado Resource Allocaton and Moblty Load Balancng n 3GPP LTE Mult-Cell Networ Feng LI,
More informationA New Quality of Service Metric for Hard/Soft Real-Time Applications
A New Qualty of Servce Metrc for Hard/Soft Real-Tme Applcatons Shaoxong Hua and Gang Qu Electrcal and Computer Engneerng Department and Insttute of Advanced Computer Study Unversty of Maryland, College
More informationMETHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS
METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng
More information