Bandwidth Allocation and Session Scheduling using SIP

Size: px
Start display at page:

Download "Bandwidth Allocation and Session Scheduling using SIP"

Transcription

1 JOURAL OF COUICAIS, VOL., O. 5, AUGUS Bandwidth Alloation and Session Sheduling using SIP Hassan HASSA, Jean-arie GARCIA and Olivier BRU LAAS-CRS, oulouse, Frane {hhassan}@laas.fr Abstrat Session Initiation Protool (SIP) is a new signaling protool designed to establish multimedia sessions in teleommuniation networks. In this paper, we suggest the extension of SIP funtionalities to oordinate QoS mehanisms deployed in IP networks, and espeially in DiffServ domain. Indeed, the interation between small and big CP sessions may have dramati onsuenes on small CP sessions. Hene, we use SIP to ahieve QoS management on a session basis, in whih the over all ativity of the user during the session is onsidered. he suggested mehanisms deal with two issues: fit, session sheduling based on session duration and/or volume, and seond bandwidth alloation on a per-flow basis using uivalent bandwidth estimation tehniques. he proposed mehanisms are implemented in the SIP proxy server as QoS management algorithms, and they are validated by simulations. Index erms SIP, DiffServ, Web, QoS, Bandwidth alloation, Sheduling. I. IRODUCI he evolution of teleommuniation networks made of the Internet a univeal platform to support most forms of modern ommuniations inluding voie, video and data appliations. However, the Internet Protool (IP) was developed based on a onnetionless model where simple metris (e.g. delay or hope ount) are used to ahieve routing. he simple onept of IP is behind its suess and its ability to sale to very large networks. Unfortunately, no Quality of Servie (QoS) was planned with this approah. Over yea, many enhanements to QoS support were implemented in IP paket networks. Hene, two QoS arhitetures DiffServ and IntServ were introdued to arry out appliation QoS ruirements. oreover, new protools suh as PLS were oneived to extend the best-effort servie of IP networks. Besides, the onvergene to all-ip network ame with new signaling protools to handle user sessions in all aess networks regardless their partiularities. hus, the Session Initiation Protool (SIP) was proposed as a signaling protool to establish and release sessions between end use. SIP is very general and an be used for any kind of sessions in all ommuniation networks. oreover in the year 000, SIP was seleted by the hird Generation Partnehip Projet (3GPP) as the all ontrol protool for the 3G IP-based mobile networks. On the other hand, the suessful deployment of ulti-label Swithing Protool PLS in DiffServ networks delegates label swithing as an apt swithing tehnology for the future ore networks espeially for its raffi Engineering (E) apabilities. Indeed, ombining both tehnologies introdues a new vision of QoS management at the appliation level. he autonomous struture of SIP makes it possible to manage user sessions as lassi phone alls, from the beginning to the end. However, SIP an ahieve muh more than signaling the beginning and the end of a ommuniation session. In partiular, it an host traffi engineering intelligene. hus, the use of SIP over DiffServ networks allows flexible QoS management as it ombines DiffServ failities with SIP supervision. In this paper, we propose a QoS management framework based on SIP over DiffServ environment. Where QoS management mehanisms are implemented and supervised by the SIP proxy server. he proposed mehanisms onerns: fit, session sheduling based on session duration and/or session data volume exhanged during a session and seond, bandwidth alloation on a per-flow basis using uivalent bandwidth estimation tehniques. he paper is organized as follows: fit, we give a brief overview of SIP and its assoiated Session Desription Protool (SDP), and then we explain the suggested SIP over DiffServ arhiteture. Seond, we present uivalent bandwidth estimation tehniques used for bandwidth alloation. Finally, we present the session based QoS algorithms implemented along with simulation results. II. SESSI IIIAI PROOCOL Session Initiation Protool is an appliation layer ontrol protool designed and developed by the IEF [4]. he speifiation is available in form of several RFCs, the most important one is RFC36 [5] whih ontains the ore protool speifiation. he easy implementation, flexibility and good salability are the main motivations onsidered while designing this protool. he main task of the protool is to set up and release sessions between end use. he session refe to the ativity between sender and reeiver when the whole state is maintained during the ommuniation. Classi sessions inlude Internet telephone alls, but it may also 006 ACADEY PUBLISHER

2 8 JOURAL OF COUICAIS, VOL., O. 5, AUGUS 006 be multimedia onferene session, Web session, distributed omputer game session, et. he ommuniation itself between devies is ahieved by other protools (often RP, RCP and SDP) as the purpose of SIP is to initiate ommuniations only. Real ime Protool RP arries the real-time appliation data (inluding audio and video) by splitting and enoding data into pakets to allow per-paket transport on the Internet, and SDP desribes and enodes apabilities of sessions. Indeed, the harateristis of the session are negotiated between partiipants. he negotiation inludes the type of odes used to enode media in order to failitate deoding proess, maximum allowed bit rates, the transport protool, et. he end-to-end model of SIP omplies with the Internet arhiteture. Indeed, all the intelligene is stored in end devies, inluding state. his protets from single point failure while preserving salability in networks. In ontrast with Publi Swithed elephone etwork (PS) where state and intelligene are stored in the network while terminals are dump. However, SIP an provide the same funtionality as PS with the possibility to implement end-to-end servies that are hardly onfigured in PS. Finally, it is lear that the salability and deentralization of SIP ome at the ost of end-to-end message overhead. In fat, SIP is based on HP protool whih is widely used on the Web. Atually, HP an be seen as a signaling protool also, as web browse tell HP serve about the douments they need. he enoding of message heade in both protools (HP and SIP) have been inherited from RFC8 [6]. his enoding has already showed robustness and flexibility with HP. he physial elements of a SIP network fall into two ategories: lients and serve. Fig., illustrates the arhiteture of a SIP network. forward SIP messages to other SIP serve in the network. Proxy server an play several roles. Besides, it provide funtions suh as authentiation, authorization, network aess ontrol, routing, reliable ruest retransmission, and seurity. Rediret server: It takes are of direting the lient to the next hop until the lient reahes the destination server and ontats UAS diretly. Registrar server: It handles lient registration ruest for its urrent loation. Generally, Registrar server is o-loated in the same physial entity hosting the rediret or proxy server. SIP works in tandem with the Session Desription Protool (SDP) that desribes multimedia sessions. Session desription serves for session announement, session invitation and other session initiation funtionalities. SDP is ompletely independent of transport protool. It onerns mainly the format of session desription and is designed to work with any transport protool. any of the SDP messages are sent using Session Announement Protool (SAP). hese messages are UDP pakets with a SAP header and a text payload. he text payload is the SDP session desription. essages an also be sent using or the World Wide Web. Fig., depits the position of SIP and SDP in the multimedia protool stak. ultimedia Appliations Audio Video Data SIP RSVP RP RCP SDP SAP Registrar Server Rediret Server CP IP UDP Figure. ultimedia protool stak. IPetwork III. SIP OVER DIFFSERV Router Router Figure. SIP Arhiteture. Client in SIP is a general onept. It ould be any devie initiating sessions (Phones, PCs, Palms, ). On the other hand, SIP serve inlude the following main types: Proxy server: he most important element in the SIP arhiteture, as it onstitutes an intermediate devie reeiving SIP ruests from lients and forwarding the ruests on lients behalf. ypially, proxy serve A. Bakground Few researhes [], [] and [3] have addressed the arhiteture of SIP over DiffServ arhiteture in IP networks. Zhang and Guy [3] proposed an extension to the Proxy server in the SIP arhiteture to inlude raffi Engineering (E) apabilities, where the SIP proxy server uses the messages exhanged during an SIP session to provide E ruests. he proxy server may use the messages exhanged during an SIP session to exhange traffi engineering ruests. hese ruests will be exhanged between the SIP proxy server and the Label Edge Router (LER) by Common Open Poliy Servie (COPS) protool messages 006 ACADEY PUBLISHER

3 JOURAL OF COUICAIS, VOL., O. 5, AUGUS []. Indeed, we need to transfer information related to the ruest of resoure by QoS lients and for the alloation of resoures by resoure alloation serve (e.g., bandwidth broker) in a DiffServ network. Hene, this resoure alloation funtionality an be added in the COPS framework. Fig. 3, depits the proposed arhiteture for SIP over DiffServ. SIP erminal E-SIP Server COPS LER PLS etwork he SIP proxy server ould only negotiate E sessions with another E enabled SIP proxy server, otherwise normal SIP session (without E extensions) is initiated. he flow of SIP messages is resumed on Fig. 4. Assuming that ommuniation issues related to resoures management and reservation at the LER is ahieved by the COPS framework; SIP an play the role of an appliation ontrol layer protool to guarantee QoS ruirements of user sessions. B. Session based QoS Generally, we speak of QoS per type of servie. hus, for real time appliations we are onerned about end to end time onstraints like delay and jitter to guarantee the reonstitution of multimedia signals (voie or video) [7]. On the other hand, non-real time appliations (or data appliations) are less sensitive to time onstraints, while they ruire guaranteed nominal bit rates and loss free transmission. Hene, these appliations (e.g Web, FP and ) use ransmission Control Protool (CP) for reliable and in-order transmission. However, the elasti behavior of CP makes the transmission ompletely dependent of the network ongestion state. In both ases, real time and non-real time appliations, the QoS ruirements are grouped together to express an SLA ruired for one appliation. his ould be onsidered as a mirosopi view of QoS. On the other hand, a marosopi view of QoS may inlude user behavior during session. he global view per session is important in our ontext beause SIP handles the session LER COPS Figure 3. SIP over DiffServ arhiteture. E-SIP Server SIP erminal SIP-E Server LER LER SIP-E Server SIP erminal Invite Invite Invite 00 ok 00 ok 00 ok COPS R COPS De raffi Dynami WFQ priorities R De SIP erminal SIP-E Server LER LER SIP-E Server SIP erminal Figure 4. SIP over DiffServ flow of messages. SIP erminal establishment and onside the user ativity during a session as a whole. he extension of SIP to offer QoS provisioning mehanisms ruires a marosopi treatment for sessions. Consuently, we define a session based QoS in whih we onsider user ativity during the session and the type of the appliation used. Based on the above presented arhiteture we want to implement QoS management algorithms at the session level. he proposed algorithms are based on traffi engineering tehniques and will be hosted in the SIP proxy server. hus, the SIP proxy server implements QoS mehanisms on multimedia sessions, based on measurements and a priori estimation of uivalent bandwidth. Indeed, the SIP proxy server measures the duration of sessions and the data volume exhanged during sessions (funtions that are generally used for billing purposes), then a session sheduling an be ahieved. Furthermore, the SIP proxy server an perform flow based uivalent bandwidth estimation, based on information olleted about session paramete before initiating it. However, this ruires uivalent bandwidth estimation tehniques that we will present in the following setion. IV. A. Related Work EQUIVALE BADWIDH raffi ontrol is generally used to optimize the alloation of network resoures in order to sustain an aeptable QoS for network onnetions. any traffi ontrol strategies inluding ongestion ontrol and all admission poliies, rely on the notion of the uivalent (or effetive) bandwidth or the resulting onnetion load on network links. For example, aess ontrol uses this information to deide whether to aept or not inoming ruests for new onnetions. he admission riteria depend on the impat of new added onnetions on both the resoure utilization and the QoS offered for aepted and already existing traffi. he notion of uivalent bandwidth has been used in the literature and two broad ategories of uivalent bandwidth estimation approahes are generally used. he fit ategory is based on Kelly s [7] mathematial definition of uivalent bandwidth for different kinds of traffi. he seond ategory refe to the analytial methods based on traditional queuing theory. he mathematial framework proposed by Kelly relies on large deviation theory to estimate the uivalent bandwidth of a stationary arrival proess. On the other hand, analytial approahes hypothesize the traffi models in order to give an approximate expression for the uivalent bandwidths in some ases suh as arkov proesses (e.g. [,0]). In this paper, we suggest an analytial approah to estimate uivalent bandwidth based on a renewal proess approximation. B. Equivalent bandwidth estimation by renewal proess approah In order to haraterize the effetive bit rate or uivalent bandwidth of a traffi soure, we need to selet an appropriate model to speify its harateristis 006 ACADEY PUBLISHER

4 0 JOURAL OF COUICAIS, VOL., O. 5, AUGUS 006 in terms of known paramete or metris. For the purposes of our researh, we adopt a two-state model (- ) that aptures the basi behavior of the data soure assoiated with a onnetion. he rationale for suh a model is that a soure is either in an idle state, ( ) transmitting at zero bit rate, or in a but state ( ) and transmitting at its peak rate. Suh a soure model has the advantage of being both simple and flexible as it an be used to either represent onnetions ranging from buty to ontinuous bit streams. Let the following values be assoiated with one onnetion: R : Average rate of one onnetion (Kbps), : Average duration of period (Se), : Average duration of period (Se), R : Average rate in period (Kbps), Q : Average file size in period (Kb) (to be used only with CP based models), X : Buffer size, E : arget paket loss probability. he renewal proess approah is an approximation of the superposition of - proesses. In this method, we study the superposition of - proesses as a GI / D // K system. In order to evaluate the uivalent bandwidth of - proesses we study the paket loss in GI / D / / K system. he paket loss probability is alulated as a funtion of the following paramete: K : he buffer size in pakets, ρ : System load with ρ= λ / µ, λ is the aggregate arrival rate of input - proesses and µ is the average servie time. a : he squared oeffiient of variation of the input arrival proess. s : he squared oeffiient of variation of the servie time. Our approah is based on paket loss approximation formulas for the GI / D // queue presented in [9,0]. Let the number of lients in the queue inluding the one being servied, be denoted by. Whitt [9] expresses the average and the seond moment of the number of lients as: ρ ( a+ s ) g E ( ) = ρ+ ( ρ) ( ) = ( )( + ) E E g is a weighing fator depending on the value of he parameter a a : () ( ρ)( ) a 3 ρ( a s ) + g e a< = () is defined as: YY = (3) Y3 Y is the value of Var ( ) given by: Var = EW VarW (4) ( ) λ ( ) ρ ρ s λ ( ) W denotes the steady state waiting time before beginning servie. he average and the variane of W are given by: τρ( a+ s ) g EW ( ) = ( ρ) VarW ( ) = E ( W ) w τ is the mean servie time and w is the squared oeffiient of variation of the waiting time expressed as: D w (5) + σ = (6) σ σ is the delay probability whose value is: σ= PW> = ρ+ ρ ρh (7) ( 0) ( a ) ( ) + a+ ρs a ( s ) (4 a s ) h + ρ + ρ + = (8) 4ρ a a+ ρ (4 a+ s ) D is the squared oeffiient of variation of the onditional delay given that the server is busy. Its value when servie time is deterministi is given by: D ρ + 4( ρ) = 3( + ) Finally Y and Y 3 are given by: s (9) ρ+ σ Y= max{ σ+ ρ, } (0) Y = max { ρ+ λew ( ), } he maximum is used to avoid division by zero. In order to ompute the paket loss when finite buffer is onsidered the two fit moments of the paket loss distribution are not suffiient. he distribution itself is needed, whih an be obtained by a ontinuous distribution fit as shown in [9]. hus Pr( > x) is expressed as a funtion of Case >.0 value as follows: γx γx Pr( > x) = pe ( e ) + + where p= p ( p) And γ=, γ= E ( ) E ( ) Case 0.99< <.0 Pr( > x) = e x E( ) 006 ACADEY PUBLISHER

5 JOURAL OF COUICAIS, VOL., O. 5, AUGUS 006 Case 0.50< < 0.99 γx γx ( γe γe ) Pr( > x) = ( γ γ) γ where γ= γe ( ) And γ= E + Var E Case < 0.50 ( ) ( ) ( ) γx Pr( > x) = e ( + γx) Where γ= E ( ) ote that the suggested heuristi estimates the paket loss probability as a funtion of buffer size, average input rate, average output rate and the squared oeffiient of variation for both arrival and servie proesses ( a and s ). Speifially, the last two paramete play an important role in estimating the uivalent bandwidth of input traffi. C. Erlang bloking probability he uivalent bandwidth estimated by the previous tehnique onside a onstant number of - onnetions. In the general ase we model the flow arrival proess at the all level as Poisson proess. Eah flow is defined by its all arrival rate λ, period average rate R, mean rate R and average duration. hus, for onnetions with i, R uivalent bandwidth, i=,...,. he overall uivalent bandwidth BW is given by: BW = Ri, () i= In the last uation we onsider onstant sessions. When Poisson arrival proess is onsidered, one an estimate the uivalent number of onnetions (iruits) for one bloking probability, using Erlang B formula: P ( ) = A /! i= 0 i A / i! () With A= λ / µ. Hene for one flow with session Poisson arrivals, we estimate the uivalent number of sessions for one bloking probability B and the uivalent bandwidth p BW is obtained easily by multiplying the uivalent rate of one onnetion by. D. Equivalent bandwidth of multimedia flows wo types of multimedia flows are onsidered: VoIP, and Data. VoIP and Data sessions are usually modeled as - proesses. However, paket transmission is ahieved by UDP for VoIP, and by CP for Data sessions. Reall that CP algorithm reats to paket losses. Hene, the uivalent bandwidth of CP flows an only be ahieved under the hypothesis of loss free transmission. However, it is suffiient in our ase as the goal is to ahieve loss free transmission. VoIP appliations have a ommon harateristi whih is the onstant paket size and onstant paket interarrival time during periods (see [5] for more details). he squared oeffiient of variation of servie time proess in a deterministi servie queue for VoIP = 0 ) beause paket sizes are pakets is null ( s onstant, while the squared oeffiient of variation of paket arrival proess for - onnetions is given by [4]: = w + w (3) a is the squared oeffiient of variation of single - onnetion and it is alulated as funtion of the paket transmission probability p, onstant paket inter-arrival, duration: p = ( / + p) (4) w= (5) + 4( ρ) ( ) he average rate is R = P * R and the maximum rate is R =. Although Web sessions share the same - struture with VoIP appliations; the paket arrival proess is more ompliated as it depends on the CP algorithm. By onsuene, the squared oeffiient of variation of paket arrival proess an not be estimated analytially. wo solutions to this problem may be proposed: fit, we an measure the value of diretly on the generated trae. his ruires having the generated traffi before evaluating the uivalent bandwidth, whih may not be useful when used in a QoS management server (the SIP proxy server). hat is why we suggest a seond heuristi based on the approximation of paket arrival proess during periods by a onstant proess of the same average. hus, we need to estimate the average rate during the period when only the file size in known. Autho in [] present a formula to alulate the transfer time when CP is used on short-lived onnetions (one ACK per two pakets b=). log.57( b) b ( ) = R ( f( prb, ) + + 4p log.57( b) + 0 p) Assuming that: (0+ 3 Rb ) + 4( pw ) W max max a (6) 006 ACADEY PUBLISHER

6 JOURAL OF COUICAIS, VOL., O. 5, AUGUS 006 b is the file size in pakets, W max he maximum reeption window, p paket loss probability, And 3.3(p+ 4p + 6 p ) + p f( pr, ) = ( + R) 0R As we are onerned with the periods of Web sessions, the file transfer ativity is very short ompared to the idle period. hus, this formula is appropriate to our study ase. eanwhile, a major simplifiation an be done when estimating the uivalent bandwidth with small loss probabilities. Indeed, the ontribution of the fit term R log.57( b ) is dominant, and the uation an be used in its simpler form: b ( ) = R log ( b) (7) Using the estimated transmission time during the period we an estimate the average transmission rate as b Q λ = with b=. b ( ) Ps Having the average transmission rate λ we an use the same formulas as for VoIP while onsidering onstant paket inter-arrivals during the period (this is an approximation). On the other hand, we onsider a maximum segment size in the CP algorithm of 984 bytes. hus, we generate pakets of (SS+Header= =04 Bytes). As a result we generate 04 bytes for all pakets exept the last one (the residual value). Consuently, the squared oeffiient of variation of servie time in a deterministi servie queue for Web pakets an be supposed null ( s 0 ). As a result, the uivalent bandwidth estimation proedure is very similar to the VoIP ase. E. Performane validation We validate the estimated uivalent bandwidth values for G7 sessions (refer to able II) and W3 Sessions (refer to able II) in network environment. For this purpose we injet the traffi generated by the two types of sessions into a queuing system of deterministi servie while the servie rate is hosen as a funtion of uivalent bandwidth estimation values. We ompare values of observed paket loss rate of both estimation methods for the uivalent bandwidth. We show results as funtion of the number of onnetions. ote that we evaluate the uivalent bandwidth with 30 paket size buffer and % paket loss rate. ABLE I..57 VALIDAI OF HE EQUIVALE BADWIDH EB VoIP Loss rate % EB Web Loss rate % Kbps Kbps he estimated uivalent bandwidth guarantees the average paket loss rate (%) espeially when the number of onnetions is important. otie that onstant paket inter-arrivals approximation during periods for Web sessions lead to aeptable results. Although this is not the real behavior of inter-arrivals when CP is used, the uivalent bandwidth estimated with this method guarantees an aeptable loss rate. his approximation allows analytial estimation of the uivalent bandwidth for Web sessions diretly. V. QOS ECHAISS WIH SIP he goal of this setion is to introdue novel mehanisms for QoS management on a per-session basis. he SIP proxy server will be delegated for the implementation of these mehanisms. he fit mehanism relies on the sheduling of CP sessions based on session duration and data volume exhanged during a session. he seond mehanism uses the uivalent bandwidth estimation methods to alloate bandwidth per flow. he SIP proxy server is supposed to ahieve measures on session durations and data volume exhanged, as well as the uivalent bandwidth estimation based on session paramete exhanged during session initiation phase. A. Sheduling of CP sessions umerous studies show that 80% of internet flows are arried by CP. It is also shown that 80-90% of the traffi is arried by only 0-0 % of the flows (big file transfe) while % of the flows arry only 0-0% of the traffi. It is obvious that CP ruires speial attention and partiularly the interation between big and small data transfe must be onsidered in any QoS provisioning mehanisms. Indeed, several researhes dealing with the effiieny of CP ongestion ontrol mehanism in ongested networks have been undertaken. However, results show that losses have dramati onsuenes on short CP onnetions. It was suggested that aording higher priority to short CP onnetions onstitutes a good solution to this problem [8]. he question of differentiating long from short CP onnetions ruires modifiations in CP heade to perform measures on CP onnetions (Duration or data volume exhanged) reader an refer to [,3] for some other proposals. In all ases this issue was always addressed at the onnetion level. On the other hand, an appliation level solution to this problem is more appropriate and easier to implement. Indeed, we an onsider the user behavior during all the session as a whole and instead of differentiating short from long onnetion we distinguish small from big sessions. Using SIP we an manage user ommuniations at the session level with session sheduling based on session level riterion. his is ahieved by supervising mehanisms implemented in the SIP Proxy server in the extended SIP arhiteture. he main advantage of our approah is that supervising mehanisms relies on measures that are performed usually for billing purposes. By onsuene, no modifiation on CP heade are ruired and no extra transmission overhead is supported. 006 ACADEY PUBLISHER

7 JOURAL OF COUICAIS, VOL., O. 5, AUGUS ) he Conept he main idea behind session sheduling is to hange traffi priority dynamially during ommuniation based on real time measurements. he goal is to minimize the impat of long CP sessions on both short CP sessions and real time traffi. Indeed, the SIP proxy server measures in real time the duration of CP sessions and data volume exhanged during one session. All sessions initiated by the SIP server has the same priority at fit. Sessions lasting more than average session duration, or exhanging more than average data volume V are automatially delassed into a lower priority lass of traffi. In order to alulate the values of and V we introdue the notion of the referene session RS. he referene session RS represents the threshold session ativity under whih the user session is onsidered as small. he notion of small session may refer to the duration or the data volume of a session. his is quiet different from the onnetion notion in whih long onnetions are synonym of big file transfe. In fat, CP sessions may ontain long idle periods and the notion of duration may lead to some wrong differentiation between sessions. One the referene session RS is defined, the theoretial duration and data volume exhanged during referene session RS an be alulated. Let: br : Average number of periods in a session, : Average duration of period (Se), : Average duration of period (Se), R : Average rate in period (Kbps), Q : Average file size in period (Kb) (to be used only with CP based models) In order to alulate the value, we onsider the CP session transmission in the free loss ase (the simplified form of CP onnetion duration uation (7)): Q Q R log.57 = Ps Ps with Ps paket size he average session duration is given by: = br ( + ) (8) And the average session data volume is given by: V = brq. (9) ) he Algorithms Here we develop the two algorithms based on referene session duration and data volume exhanged. hose algorithms are implemented by the SIP proxy server and are applied to all inoming CP sessions. Algorithm. Servie lass differentiation based on session duration/volume Calulate referene session duration and data volume V Define two servie lasses High and Low Aept all inoming CP sessions with the High servie lass For all sessions If session duration > (or session data volume > V ) End End B. Resoure alloation Delass the session servie to the Low lass he sheduling of CP sessions is a posteriori solution to network ongestion. Indeed, it minimizes interation between big and small CP session after some threshold. Although it ruires no information about CP sessions, the hoie of its threshold may problemati. Atually, it influenes the performane of the algorithm and the overall gain in terms of QoS. On the other hand, the role played by the SIP proxy server an be enhaned to suggest a priori solutions that prevent interation between big and small CP sessions. Hene, instead of deteting big CP sessions after some threshold, use may delare their sessions previously. Aording to the session type ruired by the user a different lass of servie may be assigned and onsuently an appropriate QoS is obtained. ) he Conept Resoure alloation ruires uivalent bandwidth estimation per session type. he idea is delegate the SIP proxy server to evaluate the uivalent bandwidth of CP flows per type of session. In order to ahieve this estimation the SIP proxy server needs some speifi desription of initiated sessions. his will be ahieved by the SDP protool assoiated with SIP. Assuming a Poisson arrival distribution of lient sessions, an uivalent number of sessions an be estimated by the Erlang B formula for a determined bloking probability. hen an uivalent bandwidth estimation proedure is launhed based on session information exhanged during the SDP ommuniation phase. One the uivalent bandwidth is determined, a bandwidth sharing proess is undertaken by Weighted Faire Queuing (WFQ) system at the Edge router. Partiularly, weights are hosen as funtion of uivalent bandwidth and available bandwidth. he goal is to assign the ruired bandwidth to small CP sessions, while big CP sessions share the residual bandwidth. In the following setion we will present the algorithm that will be implemented in the SIP proxy server while onsidering only two flows types. Of oue this approah may be extended to several flows of different types (not only CP sessions). ) he Algorithm Consider two types of CP sessions: small and big. he goal is to alloate resoures for small CP sessions to eliminate interation between the two types. he SIP proxy server handles the following paramete: 006 ACADEY PUBLISHER

8 4 JOURAL OF COUICAIS, VOL., O. 5, AUGUS 006 Bp he bloking probability D Poisson session arrival rate (Session/se) B Available bandwidth And the following session paramete: br : Average number of periods in a session : Average duration of period (Se) Q : Average file size in period (Kb) Algorithm : Flow based bandwidth reservation For small CP sessions alulate the uivalent number of sessions A /! Bp( ) = With A= λ i µ A / i! i= 0 Estimate the uivalent bandwidth B for sessions Define two lasses of servie S for small and B for Big Adjust the WFQ weights so that: B B B WS = and WB= B B Assign small CP sessions to the S lass of servie Assign big CP session to the B lass of servie C. Call Admission Control for CP sessions So far only the issue of minimizing interation between flows was addressed, fit by sheduling of sessions and then by resoure alloation. Indeed, a more natural role an also be assigned to the SIP proxy server, whih is Call Admission Control (CAC). Whether it onerns small or big sessions, the overhead that may be indued by the initiation of new sessions may be onsiderable when network fall into ongestion. hus, the sheduling of sessions will have no effet if the number of small CP sessions exeeds system apaity. oreover, big sessions will endure extremely long time of servie due to the system overhead. Indeed, this issue is of partiular interest for CP sessions as the session duration is tightly linked with the loss rate observed on network links. If the number of aepted sessions is larger than system apaity, higher paket loss rate may be observed and longer transmission times are needed (beause of CP retransmission mehanisms). Let use being onneting to the system aording to Poisson proess with D session arrival rate (Session/se) and let be the session duration, then the average number of use present in the system may be obtained by the Little formula: = D * (0) otie that. Hene, the inrease of session duration results in inreasing number of ative sessions in the system. herefore, the session management overhead for the SIP proxy server will inrease rapidly affeting the performane of the server itself. Call admission ontrol ould be assoiated with the previous proposed algorithms to guarantee normal funtioning of the network. his ruires the definition of the Call Admission Control hreshold ( CAC th ) per type of sessions. When resoure alloation is performed this threshold is simply defined by bandwidth reserved for one flow (denotedb ). Indeed, the estimation of B rely on the number of sessions, thus: CACth () On the other hand, when session sheduling is onsidered this is more ompliated. In fat, the result of aepting new onnetion ould not be evaluated if the uivalent bandwidth ruired by the inoming onnetion is not known. One solution may be to ahieve all admission ontrol based on real time measurements of resoure ruirements. However, in the sope of our study we will only onsider integrating CAC with resoure alloation algorithm based on bandwidth estimation. ) CAC with bandwidth alloation Basially, the session information used for uivalent bandwidth estimation and alloation is also used by CAC algorithm. hus, for every new inoming onnetion the uivalent bandwidth neessary to allow the transmission of the flow is alulated. If the estimated value does not exeed the maximum bandwidth alloated for the flow the onnetion is aepted otherwise it is rejeted. Finally, we note that resoure alloation proedure is based on session information exhanged before session initiation. he type of the session delared by the user is determinant for aepting or rejeting his demand. eanwhile, in the ase of wrong session type delaration, big sessions may be initiated as high priority ones ausing the deterioration of the overall performane of the system. herefore, it is possible to ombine the sheduling of sessions algorithms to delass wrongly delared sessions to lower lass of servie as a posteriori validation mehanism. Algorithm 3: CAC with bandwidth alloation Bandwidth alloation algorithm (same steps as in Algorithm ) For every inoming session Calulate the new uivalent session s number Estimate the uivalent bandwidth B for sessions If B < B Else End End,max Aept the onnetion Rejet the onnetion VI. SIULAIS he QoS session mehanisms proposed in previous setions are tested in a simple network of two nodes representing the two LER route. he DiffServ domain is modeled by a bottlenek link between two LER route. Bottlenek delay is of 0 ms (used for the R 006 ACADEY PUBLISHER

9 JOURAL OF COUICAIS, VOL., O. 5, AUGUS estimation). An SIP Proxy server is harged of initiating different kind sessions. Client Sessions Figure 3. We onsider three types of web sessions (W, W, W3), and a VoIP appliation session (G79). Four flows are generated (one for eah type of session) in order to evaluate the suggested QoS mehanisms. Indeed, the G79 flow is only used for performane evaluation ends while data sessions are handled dynamially by the QoS mehanisms. Session paramete are listed in able II. Session ABLE II. Q ean (Kb) Given the above session paramete we alulate the flows paramete that will be used when applying QoS mehanisms by the SIP proxy server. Bloking probability value is % for all flows (able III). ABLE III. Reall that the estimation of CP session duration is only possible in free loss transmission. his estimation is used to determine the referene duration (and volume) for sheduling of sessions. Paket transmission is ahieved by CP new Reno implemented from end to end in the simulator. Paket sizes are of 04 bytes with 40 bytes ACKS. A. est of Algorithm Q Variane (Kb) SESSI PARAEERS ean (Se) W 0KB 40KB 0 se 0 se W 50KB 00KB 0 se 40 se W3 00KB 00KB 30 se 60 se VoIP Paket Size (Bytes) E-SIP FLOW PARAEERS Flow D Session/se Se (u) LER estbed network Paket Interarrival (ms) W W W G B Kbps We onsider only two lasses of traffis High (Priority ) and Low (Priority ). All sessions start transmitting pakets in the High traffi lass. VoIP sessions stay always in the High traffi lass and do not hange their lass. Delassing sessions onerns only CP sessions. he goal of our test is to give short Web sessions represented by W sessions higher priority on other CP sessions using sheduling of sessions. In this ase, the W session is the referene session RS and its (se) G G LER E-SIP Variane (Se) (Se) transmission time without losses is the referene time ( =0.3 se ), and the data volume exhanged during the RS session is V =00 Kbytes (those values are obtained based on W session paramete). We ompare the results of proposed sheduling mehanisms with FIFO queue system without any priority lasses. Results are shown on able IV. FIFO ime Priority Volume Priority ABLE IV. Results show no big differene in performane when using time based session sheduling. Indeed, if we analyze the ativity of W sessions we note that idle periods are very long ompared to ativity time during web sessions. Delassing sessions aording to the time passed on the network is not profitable in this ase. In fat, the ommuniation duration riterion ould only be used for FP type like sessions where there are no idle times. Conveely, we see that the performane of W sessions, G79 sessions has improved onsiderably in the ase of session delassing based on data volume exhanged during a session. In fat, W session duration is loser to theoretial value without losses, while G79 sessions endure less paket loss rate. B. est of algorithm Using the same previous example sessions, we alulate the weights of WFQ system based on the estimation of the uivalent bandwidth per flow. We onsider two WFQ queues in the edge router with two orresponding traffi lasses: High and Low. he weights are alulated to alloate the ruired bandwidth for the High traffi lass. he Low traffi lass takes the residual bandwidth. able V depits the alulated weights based on sessions paramete defined before. ABLE V. Flow B Kbps FLOWS SAISICS WIH DIFFERE PRIORIIES Flow Loss % Delay ms Average session duration se W 3% W.9% 5 40 W3.7% G79 3.5% W.9% 5 70 W.5% W3.3% G79.6% W 0.9% W.7% W3 3.% G79.% WFQ WEIGHS FOR BADWIDH SHARIG raffi lass Weight Bottlenek bandwidth kbps W+G79C 43.7 High 00 W+W Low ACADEY PUBLISHER

10 6 JOURAL OF COUICAIS, VOL., O. 5, AUGUS 006 Bandwidth alloation Results are better than fit mehanism. Loss rate and delay on the W flow and G79 flow are smaller than previous tests. Espeially, W session average duration is very lose to theoretial estimation. Even W and W3 sessions performs better ompared to the FIFO system ase. his is espeially due to the better bandwidth utilization and the isolation fator resulting in less interation between flows. o illustrate the robustness of this approah we show on Fig. 5 the evolution of session durations while the link bandwidth is redued from 00 Kbps to 800 Kbps by 00 Kbps step. he SIP proxy server adjusts weights of the WFQ queuing system to guarantee the ruired bandwidth for W and G79 flows while the W and W3 flows gets always the residual bandwidth. he urves show that the measured W session duration is stable while the measured W and W3 session durations inreases as the bandwidth is redued. Session Duration in Se ABLE VI FLOWS SAISICS WIH WFQ Flow Loss % Delay ms Average session duration se W 0.3% W % W3.% G % Evolution of session duration vs bottlenek Bottlenek in Kbps Figure 5. Evolution of session s durations vs bottlenek he stability of the duration of small CP session ould be seen as a QoS parameter for Web sessions always served as best effort traffi. It has bigger impat in wireless networks where bandwidth is a preious resoure, and guaranteed average throughput is an important parameter of the servie. VII. CCLUSI In this paper, we presented an SIP based framework for QoS at the session level. Using the SIP proxy server with extended arhiteture over DiffServ domain, we an implement session sheduling and bandwidth alloation mehanisms to minimize the interation between small and big CP sessions. he suggested mehanisms rely on session level real time measurements of session duration and data volume exhanged during a session as well as W W W3 uivalent bandwidth estimation tehnique. he ruired measurements do not ause additional overhead as they are ahieved for billing purposes. However, bandwidth alloation ruires more speifi session desription that should be exhanged between the SIP proxy and the lient using the SDP protool. Bandwidth alloation is based on uivalent bandwidth estimation per flow. Indeed, we proposed a GI / D / / K queue system model to evaluate the uivalent bandwidth using the fit and seond order moment of paket inter-arrivals and paket servie proesses. However, some heuristis to measure the uivalent bandwidth in real time ould be onsidered using iterative algorithms. his may be useful also to implement adaptive all admission ontrol mehanisms. During our study we only onsidered homogenous flows (per type of session) for the estimation of uivalent bandwidth. In fat, the estimated uivalent bandwidth depends on traffi mixes and may be influened by the different paket sizes aording to appliation types (and transport protools). Partiularly, in VoIP appliation paket sizes are very small omparing to Video and CP sessions resulting in important ovariane on the paket servie proess and the overall performane. Finally, we would like to note the QoS mehanisms proposed are applied on a per-session basis, and this may result in an extra delay on session initiation proess for use. Some aggregated reservation tehniques [] may be useful in this ase to enhane the SIP proxy server response time. REFERECES [] F. P. Kelly. Effetive bandwidths at multi-lass queues. Queueing Systems, 9:5--6,99. [] R. Geurin, H. Ahmadi, and. aghshineh. Equivalent apaity and its appliation to bandwidth alloation in high-speed networks. IEEE J. Selet. Areas Commun, 9(7):968-98, September 99. [3] C. Couroubetis and R.Weber. Effetive bandwidths for stationary soures. Prob. Eng. Inf. Si., 9:85--96, 995. [4] A. Elwalid and D. itra. Effetive bandwidth of general arkovian traffi soures and admission ontrol of high speed networks. IEEE/AC rans. on etworking, (3): , Otober 993. [5] C. Couroubetis and R. Weber. Buffer overflow asymptotis for a swith handling many traffi soures. Journal of Applied Probability, 33: , 996. [6] A. Simonian and J. Guibert. Large deviations approximations for fluid queues fed by a large number of on-off soures. IEEE J. Selet. Areas Commun., 3(7):07--07, August 995. [7] F. P. Kelly. otes on effetive bandwidths. In F. P. Kelly, S. Zahary, and I. Zeidins, edito, Stohasti etworks: heory and Appliations, pages Oxford Univeity Press, 996. [8] G. de Veiana and J. Walrand. Effetive bandwidths: Call admission, traffi poliing and filtering in A networks. Queuing Systems, 0:37--59, 995. [9]. Likhanov and R. R. azumdar. Cell loss asymptotis for buffe fed with a large number of independent 006 ACADEY PUBLISHER

11 JOURAL OF COUICAIS, VOL., O. 5, AUGUS stationary soures. Pro. of IEEE IFOCO'98, San Franiso, USA, arh 998. [0] W. Whitt, ail probabilities with statistial multiplexing and effetive bandwidths in multi-lass queues, eleommuniation Systems, pp.7-07, 993. [] B. Rong, J. Lebeau,. Bennani,. Kadoh and A. Elhakeem, raffi Aggregation Based SIP over PLS etwork Arhiteture, Proeedings of International Conferene on Advaned Information etworking and Appliations (AIA 005), 005. [] S. Salsano and S. Veltri, QoS Control by eans of COPS to Support SIP-Based Appliations, IEEE etwork, Vol. 6, Issue :, pages 7-33, 00. [3] C. Zhang and C.G Guy, E-SIP Server Design for a SIPover-PLS Based etwork, Proeedings of International Conferene on Communiation ehnology (ICC 003), Volume :, pages , 003. [4] [5] [6] [7] La qualité de servie de la voix sur IP, prinipes et assurane. [8] U. Ayesta, Stohasti Sheduling and its Appliation to CP/IP etworks, hesis of IRIA, ie-sophia Antipolie, 005. [9] W. Whitt, he queuing network analyzer. Bell System ehnial Journal, pages , ovember 983. [0] R. agarajan, J. F. Kurose, D. owsley, Approximation ehniques for Computing Paket Loss in Finite-Buffered Voie ultiplexe, IEEE Journal on Seleted Areas in Communiations (99). [] B Sikdar, S Kalyanaramana and K.S Vastola, An Integrated odel for the Lateny and Steady-State hroughput of CP onnetions. Performane Evaluation, Vol 46, no. -3, pp , 00. [] C. Estan and G. Varghese, ew diretions in traffi measurement and aounting: Fousing on the Elephants, ignoring the ie AC transations on Computer Systems, August 003. [3] C. Psounis A. Ghosh and B. Prabhakar. SIF: a simple algorithm for identifying large flows, Presented at the Stohasti etworks Conferene, ontreal, Canada, 004. [4] K. Sriram and W. Whitt, Charaterizing Superposition Arrival Proesses in Paket ultiplexe for Voie and Data, IEEE Journal of Seleted Areas in Communiations, pp , Sep, 986. [5] H. Hassan, J- Garia and C. Bokstal Aggregate traffi models for VoIP appliations, ICD 006, in press. Hassan HASSA. is a Ph.D. student at Laboratoire d Arhiteture et d Analyse de Systèmes of CRS, oulouse, Frane. He reeived his S degree in etworks and eleommuniations from the ESEEIH, oulouse, Frane, in 003 and his B.E. degree in Eletronis from ESERG, Grenoble, Frane, in 996. His researh interests inlude ultimedia raffi odeling and Performane Analysis in Heterogeneous etworks. Jean-arie GARCIA. reeived his Ph.D. from the Univeité Paul Sabatier, oulouse, Frane, in 980. He reeived his B.E. degree from Institut ational des Sienes Appliquées, oulouse, Frane, in 976, and the. A. S degree from the Eole Polytehnique, ontreal, P.Q., Canada, in 978. He performed Researh at the Institut of Hydro Quebe, Varennes, P.Q., Canada, in 977. He spent one year as a postdotoral fellow at the Eletronis Researh Laboratory, Univeity of California, Berkeley, in 98, and in 98 he joind the Laboratoire d'analyse et d'arhiteture de Systèmes of CRS, oulouse, where he is presently direteur de reherhes. His main interests are modeling and ontrol of teleommuniation networks, optimization and resoure management aspets. Olivier BRU. reeived his Ph.D. from the Univeité Paul Sabatier, oulouse, Frane, in 000. He reeived his B.E. degree from Institut ational des eleommuniations, Evry, Frane, in 995. He spent one year as a researh engineer in the Delta Partne ompany where he developed a modeling tool for large-sale telephone networks. In 00, he joined the Laboratoire d'analyse et d'arhiteture de Systèmes of CRS, oulouse, where he is presently harge de reherhes. His researh deals with stohasti modeling and network planning. 006 ACADEY PUBLISHER

Performance Analysis of IEEE 802.11 in Multi-hop Wireless Networks

Performance Analysis of IEEE 802.11 in Multi-hop Wireless Networks Performane Analysis of IEEE 80.11 in Multi-hop Wireless Networks Lan Tien Nguyen 1, Razvan Beuran,1, Yoihi Shinoda 1, 1 Japan Advaned Institute of Siene and Tehnology, 1-1 Asahidai, Nomi, Ishikawa, 93-19

More information

A Holistic Method for Selecting Web Services in Design of Composite Applications

A Holistic Method for Selecting Web Services in Design of Composite Applications A Holisti Method for Seleting Web Servies in Design of Composite Appliations Mārtiņš Bonders, Jānis Grabis Institute of Information Tehnology, Riga Tehnial University, 1 Kalu Street, Riga, LV 1658, Latvia,

More information

Henley Business School at Univ of Reading. Pre-Experience Postgraduate Programmes Chartered Institute of Personnel and Development (CIPD)

Henley Business School at Univ of Reading. Pre-Experience Postgraduate Programmes Chartered Institute of Personnel and Development (CIPD) MS in International Human Resoure Management For students entering in 2012/3 Awarding Institution: Teahing Institution: Relevant QAA subjet Benhmarking group(s): Faulty: Programme length: Date of speifiation:

More information

Static Fairness Criteria in Telecommunications

Static Fairness Criteria in Telecommunications Teknillinen Korkeakoulu ERIKOISTYÖ Teknillisen fysiikan koulutusohjelma 92002 Mat-208 Sovelletun matematiikan erikoistyöt Stati Fairness Criteria in Teleommuniations Vesa Timonen, e-mail: vesatimonen@hutfi

More information

Hierarchical Clustering and Sampling Techniques for Network Monitoring

Hierarchical Clustering and Sampling Techniques for Network Monitoring S. Sindhuja Hierarhial Clustering and Sampling Tehniques for etwork Monitoring S. Sindhuja ME ABSTRACT: etwork monitoring appliations are used to monitor network traffi flows. Clustering tehniques are

More information

Supply chain coordination; A Game Theory approach

Supply chain coordination; A Game Theory approach aepted for publiation in the journal "Engineering Appliations of Artifiial Intelligene" 2008 upply hain oordination; A Game Theory approah Jean-Claude Hennet x and Yasemin Arda xx x LI CNR-UMR 668 Université

More information

Deadline-based Escalation in Process-Aware Information Systems

Deadline-based Escalation in Process-Aware Information Systems Deadline-based Esalation in Proess-Aware Information Systems Wil M.P. van der Aalst 1,2, Mihael Rosemann 2, Marlon Dumas 2 1 Department of Tehnology Management Eindhoven University of Tehnology, The Netherlands

More information

Computer Networks Framing

Computer Networks Framing Computer Networks Framing Saad Mneimneh Computer Siene Hunter College of CUNY New York Introdution Who framed Roger rabbit? A detetive, a woman, and a rabbit in a network of trouble We will skip the physial

More information

Capacity at Unsignalized Two-Stage Priority Intersections

Capacity at Unsignalized Two-Stage Priority Intersections Capaity at Unsignalized Two-Stage Priority Intersetions by Werner Brilon and Ning Wu Abstrat The subjet of this paper is the apaity of minor-street traffi movements aross major divided four-lane roadways

More information

OpenScape 4000 CSTA V7 Connectivity Adapter - CSTA III, Part 2, Version 4.1. Developer s Guide A31003-G9310-I200-1-76D1

OpenScape 4000 CSTA V7 Connectivity Adapter - CSTA III, Part 2, Version 4.1. Developer s Guide A31003-G9310-I200-1-76D1 OpenSape 4000 CSTA V7 Connetivity Adapter - CSTA III, Part 2, Version 4.1 Developer s Guide A31003-G9310-I200-1-76 Our Quality and Environmental Management Systems are implemented aording to the requirements

More information

MATE: MPLS Adaptive Traffic Engineering

MATE: MPLS Adaptive Traffic Engineering MATE: MPLS Adaptive Traffi Engineering Anwar Elwalid Cheng Jin Steven Low Indra Widjaja Bell Labs EECS Dept EE Dept Fujitsu Network Communiations Luent Tehnologies Univ. of Mihigan Calteh Pearl River,

More information

Learning Curves and Stochastic Models for Pricing and Provisioning Cloud Computing Services

Learning Curves and Stochastic Models for Pricing and Provisioning Cloud Computing Services T Learning Curves and Stohasti Models for Priing and Provisioning Cloud Computing Servies Amit Gera, Cathy H. Xia Dept. of Integrated Systems Engineering Ohio State University, Columbus, OH 4310 {gera.,

More information

Henley Business School at Univ of Reading. Chartered Institute of Personnel and Development (CIPD)

Henley Business School at Univ of Reading. Chartered Institute of Personnel and Development (CIPD) MS in International Human Resoure Management (full-time) For students entering in 2015/6 Awarding Institution: Teahing Institution: Relevant QAA subjet Benhmarking group(s): Faulty: Programme length: Date

More information

State of Maryland Participation Agreement for Pre-Tax and Roth Retirement Savings Accounts

State of Maryland Participation Agreement for Pre-Tax and Roth Retirement Savings Accounts State of Maryland Partiipation Agreement for Pre-Tax and Roth Retirement Savings Aounts DC-4531 (08/2015) For help, please all 1-800-966-6355 www.marylandd.om 1 Things to Remember Complete all of the setions

More information

Dataflow Features in Computer Networks

Dataflow Features in Computer Networks Magyar Kutatók 1. Nemzetközi Szimpóziuma 1 th International Symposium of Hungarian Researhers on Computational Intelligene and Informatis Gyula Max Budapest University of Tehnology and Eonomis, Department

More information

Sebastián Bravo López

Sebastián Bravo López Transfinite Turing mahines Sebastián Bravo López 1 Introdution With the rise of omputers with high omputational power the idea of developing more powerful models of omputation has appeared. Suppose that

More information

Table of Contents. Appendix II Application Checklist. Export Finance Program Working Capital Financing...7

Table of Contents. Appendix II Application Checklist. Export Finance Program Working Capital Financing...7 Export Finane Program Guidelines Table of Contents Setion I General...........................................................1 A. Introdution............................................................1

More information

Parametric model of IP-networks in the form of colored Petri net

Parametric model of IP-networks in the form of colored Petri net Parametri model of IP-networks in the form of olored Petri net Shmeleva T.R. Abstrat A parametri model of IP-networks in the form of olored Petri net was developed; it onsists of a fixed number of Petri

More information

Unit 12: Installing, Configuring and Administering Microsoft Server

Unit 12: Installing, Configuring and Administering Microsoft Server Unit 12: Installing, Configuring and Administering Mirosoft Server Learning Outomes A andidate following a programme of learning leading to this unit will be able to: Selet a suitable NOS to install for

More information

Granular Problem Solving and Software Engineering

Granular Problem Solving and Software Engineering Granular Problem Solving and Software Engineering Haibin Zhu, Senior Member, IEEE Department of Computer Siene and Mathematis, Nipissing University, 100 College Drive, North Bay, Ontario, P1B 8L7, Canada

More information

AUDITING COST OVERRUN CLAIMS *

AUDITING COST OVERRUN CLAIMS * AUDITING COST OVERRUN CLAIMS * David Pérez-Castrillo # University of Copenhagen & Universitat Autònoma de Barelona Niolas Riedinger ENSAE, Paris Abstrat: We onsider a ost-reimbursement or a ost-sharing

More information

Agent-Based Grid Load Balancing Using Performance-Driven Task Scheduling

Agent-Based Grid Load Balancing Using Performance-Driven Task Scheduling Agent-Based Grid Load Balaning Using Performane-Driven Task Sheduling Junwei Cao *1, Daniel P. Spooner, Stephen A. Jarvis, Subhash Saini and Graham R. Nudd * C&C Researh Laboratories, NEC Europe Ltd.,

More information

Open and Extensible Business Process Simulator

Open and Extensible Business Process Simulator UNIVERSITY OF TARTU FACULTY OF MATHEMATICS AND COMPUTER SCIENCE Institute of Computer Siene Karl Blum Open and Extensible Business Proess Simulator Master Thesis (30 EAP) Supervisors: Luiano Garía-Bañuelos,

More information

PROCEEDS OF CRIME (BUSINESS IN THE REGULATED SECTOR) ORDER 2015

PROCEEDS OF CRIME (BUSINESS IN THE REGULATED SECTOR) ORDER 2015 Proeeds of Crime (Business in the Regulated Setor) Order 2015 Artile 1 Statutory Doument No. 2015/0073 Proeeds of Crime At 2008 PROCEEDS OF CRIME (BUSINESS IN THE REGULATED SECTOR) ORDER 2015 Approved

More information

Scalable TCP: Improving Performance in Highspeed Wide Area Networks

Scalable TCP: Improving Performance in Highspeed Wide Area Networks Salable TP: Improving Performane in Highspeed Wide Area Networks Tom Kelly ERN - IT Division Geneva 3 Switzerland tk@am.a.uk ABSTRAT TP ongestion ontrol an perform badly in highspeed wide area networks

More information

IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, VOL. 9, NO. 3, MAY/JUNE 2012 401

IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, VOL. 9, NO. 3, MAY/JUNE 2012 401 IEEE TRASACTIOS O DEPEDABLE AD SECURE COMPUTIG, VOL. 9, O. 3, MAY/JUE 2012 401 Mitigating Distributed Denial of Servie Attaks in Multiparty Appliations in the Presene of Clok Drifts Zhang Fu, Marina Papatriantafilou,

More information

Using Live Chat in your Call Centre

Using Live Chat in your Call Centre Using Live Chat in your Call Centre Otober Key Highlights Yesterday's all entres have beome today's ontat entres where agents deal with multiple queries from multiple hannels. Live Chat hat is one now

More information

INTELLIGENCE IN SWITCHED AND PACKET NETWORKS

INTELLIGENCE IN SWITCHED AND PACKET NETWORKS 22-24 Setember, Sozool, ULGRI INTELLIGENCE IN SWITCHED ND PCKET NETWORKS Ivaylo Ivanov tanasov Deartment of teleommuniations, Tehnial University of Sofia, 7 Kliment Ohridski st., 1000, hone: +359 2 965

More information

Customer Reporting for SaaS Applications. Domain Basics. Managing my Domain

Customer Reporting for SaaS Applications. Domain Basics. Managing my Domain Produtivity Marketpla e Software as a Servie Invoiing Ordering Domains Customer Reporting for SaaS Appliations Domain Basis Managing my Domain Managing Domains Helpful Resoures Managing my Domain If you

More information

Green Cloud Computing

Green Cloud Computing International Journal of Information and Computation Tehnology. ISSN 0974-2239 Volume 4, Number 4 (2014), pp. 431-436 International Researh Publiations House http://www. irphouse.om /ijit.htm Green Cloud

More information

A Keyword Filters Method for Spam via Maximum Independent Sets

A Keyword Filters Method for Spam via Maximum Independent Sets Vol. 7, No. 3, May, 213 A Keyword Filters Method for Spam via Maximum Independent Sets HaiLong Wang 1, FanJun Meng 1, HaiPeng Jia 2, JinHong Cheng 3 and Jiong Xie 3 1 Inner Mongolia Normal University 2

More information

Neural network-based Load Balancing and Reactive Power Control by Static VAR Compensator

Neural network-based Load Balancing and Reactive Power Control by Static VAR Compensator nternational Journal of Computer and Eletrial Engineering, Vol. 1, No. 1, April 2009 Neural network-based Load Balaning and Reative Power Control by Stati VAR Compensator smail K. Said and Marouf Pirouti

More information

Chapter 1 Microeconomics of Consumer Theory

Chapter 1 Microeconomics of Consumer Theory Chapter 1 Miroeonomis of Consumer Theory The two broad ategories of deision-makers in an eonomy are onsumers and firms. Eah individual in eah of these groups makes its deisions in order to ahieve some

More information

An Enhanced Critical Path Method for Multiple Resource Constraints

An Enhanced Critical Path Method for Multiple Resource Constraints An Enhaned Critial Path Method for Multiple Resoure Constraints Chang-Pin Lin, Hung-Lin Tai, and Shih-Yan Hu Abstrat Traditional Critial Path Method onsiders only logial dependenies between related ativities

More information

From a strategic view to an engineering view in a digital enterprise

From a strategic view to an engineering view in a digital enterprise Digital Enterprise Design & Management 2013 February 11-12, 2013 Paris From a strategi view to an engineering view in a digital enterprise The ase of a multi-ountry Telo Hervé Paault Orange Abstrat In

More information

AT 6 OF 2012 GAMBLING DUTY ACT 2012

AT 6 OF 2012 GAMBLING DUTY ACT 2012 i e AT 6 OF 2012 GAMBLING DUTY ACT 2012 Gambling Duty At 2012 Index i e GAMBLING DUTY ACT 2012 Index Setion Page PART 1 INTRODUCTORY 5 1 Short title... 5 2 Commenement... 5 3 General interpretation...

More information

protection p1ann1ng report

protection p1ann1ng report f1re~~ protetion p1ann1ng report BUILDING CONSTRUCTION INFORMATION FROM THE CONCRETE AND MASONRY INDUSTRIES Signifiane of Fire Ratings for Building Constrution NO. 3 OF A SERIES The use of fire-resistive

More information

Channel Assignment Strategies for Cellular Phone Systems

Channel Assignment Strategies for Cellular Phone Systems Channel Assignment Strategies for Cellular Phone Systems Wei Liu Yiping Han Hang Yu Zhejiang University Hangzhou, P. R. China Contat: wliu5@ie.uhk.edu.hk 000 Mathematial Contest in Modeling (MCM) Meritorious

More information

An integrated optimization model of a Closed- Loop Supply Chain under uncertainty

An integrated optimization model of a Closed- Loop Supply Chain under uncertainty ISSN 1816-6075 (Print), 1818-0523 (Online) Journal of System and Management Sienes Vol. 2 (2012) No. 3, pp. 9-17 An integrated optimization model of a Closed- Loop Supply Chain under unertainty Xiaoxia

More information

Deduplication with Block-Level Content-Aware Chunking for Solid State Drives (SSDs)

Deduplication with Block-Level Content-Aware Chunking for Solid State Drives (SSDs) 23 IEEE International Conferene on High Performane Computing and Communiations & 23 IEEE International Conferene on Embedded and Ubiquitous Computing Dedupliation with Blok-Level Content-Aware Chunking

More information

Customer Efficiency, Channel Usage and Firm Performance in Retail Banking

Customer Efficiency, Channel Usage and Firm Performance in Retail Banking Customer Effiieny, Channel Usage and Firm Performane in Retail Banking Mei Xue Operations and Strategi Management Department The Wallae E. Carroll Shool of Management Boston College 350 Fulton Hall, 140

More information

SLA-based Resource Allocation for Software as a Service Provider (SaaS) in Cloud Computing Environments

SLA-based Resource Allocation for Software as a Service Provider (SaaS) in Cloud Computing Environments 2 th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing SLA-based Resoure Alloation for Software as a Servie Provider (SaaS) in Cloud Computing Environments Linlin Wu, Saurabh Kumar

More information

A Comparison of Service Quality between Private and Public Hospitals in Thailand

A Comparison of Service Quality between Private and Public Hospitals in Thailand International Journal of Business and Soial Siene Vol. 4 No. 11; September 2013 A Comparison of Servie Quality between Private and Hospitals in Thailand Khanhitpol Yousapronpaiboon, D.B.A. Assistant Professor

More information

An Efficient Network Traffic Classification Based on Unknown and Anomaly Flow Detection Mechanism

An Efficient Network Traffic Classification Based on Unknown and Anomaly Flow Detection Mechanism An Effiient Network Traffi Classifiation Based on Unknown and Anomaly Flow Detetion Mehanism G.Suganya.M.s.,B.Ed 1 1 Mphil.Sholar, Department of Computer Siene, KG College of Arts and Siene,Coimbatore.

More information

' R ATIONAL. :::~i:. :'.:::::: RETENTION ':: Compliance with the way you work PRODUCT BRIEF

' R ATIONAL. :::~i:. :'.:::::: RETENTION ':: Compliance with the way you work PRODUCT BRIEF ' R :::i:. ATIONAL :'.:::::: RETENTION ':: Compliane with the way you work, PRODUCT BRIEF In-plae Management of Unstrutured Data The explosion of unstrutured data ombined with new laws and regulations

More information

WORKFLOW CONTROL-FLOW PATTERNS A Revised View

WORKFLOW CONTROL-FLOW PATTERNS A Revised View WORKFLOW CONTROL-FLOW PATTERNS A Revised View Nik Russell 1, Arthur H.M. ter Hofstede 1, 1 BPM Group, Queensland University of Tehnology GPO Box 2434, Brisbane QLD 4001, Australia {n.russell,a.terhofstede}@qut.edu.au

More information

CARRIER VOICE NETWORK MANAGEMENT

CARRIER VOICE NETWORK MANAGEMENT CARRIER VOICE NETWORK MANAGEMENT Elvis N. Ngah Supervised by: Dr. Sandjai Bhulai Department of Mathematis Vrije Universiteit Amsterdam The Netherlands Deryk Cole Network Operations Center Telesonique Global

More information

Impedance Method for Leak Detection in Zigzag Pipelines

Impedance Method for Leak Detection in Zigzag Pipelines 10.478/v10048-010-0036-0 MEASUREMENT SCIENCE REVIEW, Volume 10, No. 6, 010 Impedane Method for Leak Detetion in igzag Pipelines A. Lay-Ekuakille 1, P. Vergallo 1, A. Trotta 1 Dipartimento d Ingegneria

More information

SOFTWARE ENGINEERING I

SOFTWARE ENGINEERING I SOFTWARE ENGINEERING I CS 10 Catalog Desription PREREQUISITE: CS 21. Introdution to the systems development life yle, software development models, analysis and design tehniques and tools, and validation

More information

TECHNOLOGY-ENHANCED LEARNING FOR MUSIC WITH I-MAESTRO FRAMEWORK AND TOOLS

TECHNOLOGY-ENHANCED LEARNING FOR MUSIC WITH I-MAESTRO FRAMEWORK AND TOOLS TECHNOLOGY-ENHANCED LEARNING FOR MUSIC WITH I-MAESTRO FRAMEWORK AND TOOLS ICSRiM - University of Leeds Shool of Computing & Shool of Musi Leeds LS2 9JT, UK +44-113-343-2583 kia@i-maestro.org www.i-maestro.org,

More information

A Context-Aware Preference Database System

A Context-Aware Preference Database System J. PERVASIVE COMPUT. & COMM. (), MARCH 005. TROUBADOR PUBLISHING LTD) A Context-Aware Preferene Database System Kostas Stefanidis Department of Computer Siene, University of Ioannina,, kstef@s.uoi.gr Evaggelia

More information

Course 8-9 DSL type digital access techniques (Digital Subscriber Line)

Course 8-9 DSL type digital access techniques (Digital Subscriber Line) Course 8-9 DSL type digital aess tehniques (Digital Subsriber Line) o The term refers to the tehnologies and equipments used in a telephone network to ensure the aess to a high speed digital network on

More information

Strategies for Development and Adoption of ERR in German Ambulatory Care

Strategies for Development and Adoption of ERR in German Ambulatory Care Strategies for Development and Adoption of ERR in German Ambulatory Care Sebastian Duennebeil 1, Ali Sunyaev 1, Jan Maro Leimeister 2, Helmut Krmar 1 1 Department of Informatis 1 Tehnishe Universitat MUnhen

More information

A Game Theoretical Approach to Gateway Selections in Multi-domain Wireless Networks

A Game Theoretical Approach to Gateway Selections in Multi-domain Wireless Networks 1 A Game Theoretial Approah to Gateway Seletions in Multi-domain Wireless Networks Yang Song, Starsky H.Y. Wong and Kang-Won Lee IBM Researh, Hawthorne, NY Email: {yangsong, hwong, kangwon}@us.ibm.om Abstrat

More information

OpenSession: SDN-based Cross-layer Multi-stream Management Protocol for 3D Teleimmersion

OpenSession: SDN-based Cross-layer Multi-stream Management Protocol for 3D Teleimmersion : SN-based ross-layer Multi-stream Management Protool for 3 Teleimmersion hsan refin Raoul Rivas Rehana Tabassum Klara Nahrstedt University of Illinois at Urbana-hampaign, Urbana, IL {marefin2, trivas,

More information

Improved SOM-Based High-Dimensional Data Visualization Algorithm

Improved SOM-Based High-Dimensional Data Visualization Algorithm Computer and Information Siene; Vol. 5, No. 4; 2012 ISSN 1913-8989 E-ISSN 1913-8997 Published by Canadian Center of Siene and Eduation Improved SOM-Based High-Dimensional Data Visualization Algorithm Wang

More information

A Reputation Management Approach for Resource Constrained Trustee Agents

A Reputation Management Approach for Resource Constrained Trustee Agents A Reputation Management Approah for Resoure Constrained rustee Agents Han Yu, Chunyan Miao, Bo An 2, Cyril Leung 3, and Vitor R. Lesser 4 Nanyang ehnologial University, Singapore 2 he Key Lab of Intelligent

More information

i_~f e 1 then e 2 else e 3

i_~f e 1 then e 2 else e 3 A PROCEDURE MECHANISM FOR BACKTRACK PROGRAMMING* David R. HANSON + Department o Computer Siene, The University of Arizona Tuson, Arizona 85721 One of the diffiulties in using nondeterministi algorithms

More information

FIRE DETECTION USING AUTONOMOUS AERIAL VEHICLES WITH INFRARED AND VISUAL CAMERAS. J. Ramiro Martínez-de Dios, Luis Merino and Aníbal Ollero

FIRE DETECTION USING AUTONOMOUS AERIAL VEHICLES WITH INFRARED AND VISUAL CAMERAS. J. Ramiro Martínez-de Dios, Luis Merino and Aníbal Ollero FE DETECTION USING AUTONOMOUS AERIAL VEHICLES WITH INFRARED AND VISUAL CAMERAS. J. Ramiro Martínez-de Dios, Luis Merino and Aníbal Ollero Robotis, Computer Vision and Intelligent Control Group. University

More information

Board Building Recruiting and Developing Effective Board Members for Not-for-Profit Organizations

Board Building Recruiting and Developing Effective Board Members for Not-for-Profit Organizations Board Development Board Building Reruiting and Developing Effetive Board Members for Not-for-Profit Organizations Board Development Board Building Reruiting and Developing Effetive Board Members for Not-for-Profit

More information

INCOME TAX WITHHOLDING GUIDE FOR EMPLOYERS

INCOME TAX WITHHOLDING GUIDE FOR EMPLOYERS Virginia Department of Taxation INCOME TAX WITHHOLDING GUIDE FOR EMPLOYERS www.tax.virginia.gov 2614086 Rev. 07/14 * Table of Contents Introdution... 1 Important... 1 Where to Get Assistane... 1 Online

More information

On the Characteristics of Spectrum-Agile Communication Networks

On the Characteristics of Spectrum-Agile Communication Networks 1 On the Charateristis of Spetrum-Agile Communiation Networks Xin Liu Wei Wang Department of Computer Siene University of California, Davis, CA 95616 Email:{liu,wangw}@s.udavis.edu Abstrat Preliminary

More information

Price-based versus quantity-based approaches for stimulating the development of renewable electricity: new insights in an old debate

Price-based versus quantity-based approaches for stimulating the development of renewable electricity: new insights in an old debate Prie-based versus -based approahes for stimulating the development of renewable eletriity: new insights in an old debate uthors: Dominique FINON, Philippe MENNTEU, Marie-Laure LMY, Institut d Eonomie et

More information

Optimal Allocation of Filters against DDoS Attacks

Optimal Allocation of Filters against DDoS Attacks Optimal Alloation of Filters against DDoS Attaks Karim El Defrawy ICS Dept. University of California, Irvine Email: keldefra@ui.edu Athina Markopoulou EECS Dept. University of California, Irvine Email:

More information

A novel active mass damper for vibration control of bridges

A novel active mass damper for vibration control of bridges IABMAS 08, International Conferene on Bridge Maintenane, Safety and Management, 3-7 July 008, Seoul, Korea A novel ative mass damper for vibration ontrol of bridges U. Starossek & J. Sheller Strutural

More information

IT Essentials II: Network Operating Systems

IT Essentials II: Network Operating Systems Unit 13: IT Essentials II: Network Operating Systems Learning Outomes: A andidate following a programme of learning leading to this unit will be able to: demonstrate understanding of network operating

More information

Weighting Methods in Survey Sampling

Weighting Methods in Survey Sampling Setion on Survey Researh Methods JSM 01 Weighting Methods in Survey Sampling Chiao-hih Chang Ferry Butar Butar Abstrat It is said that a well-designed survey an best prevent nonresponse. However, no matter

More information

Information Security 201

Information Security 201 FAS Information Seurity 201 Desktop Referene Guide Introdution Harvard University is ommitted to proteting information resoures that are ritial to its aademi and researh mission. Harvard is equally ommitted

More information

Computational Analysis of Two Arrangements of a Central Ground-Source Heat Pump System for Residential Buildings

Computational Analysis of Two Arrangements of a Central Ground-Source Heat Pump System for Residential Buildings Computational Analysis of Two Arrangements of a Central Ground-Soure Heat Pump System for Residential Buildings Abstrat Ehab Foda, Ala Hasan, Kai Sirén Helsinki University of Tehnology, HVAC Tehnology,

More information

THE PERFORMANCE OF TRANSIT TIME FLOWMETERS IN HEATED GAS MIXTURES

THE PERFORMANCE OF TRANSIT TIME FLOWMETERS IN HEATED GAS MIXTURES Proeedings of FEDSM 98 998 ASME Fluids Engineering Division Summer Meeting June 2-25, 998 Washington DC FEDSM98-529 THE PERFORMANCE OF TRANSIT TIME FLOWMETERS IN HEATED GAS MIXTURES John D. Wright Proess

More information

Electrician'sMathand BasicElectricalFormulas

Electrician'sMathand BasicElectricalFormulas Eletriian'sMathand BasiEletrialFormulas MikeHoltEnterprises,In. 1.888.NEC.CODE www.mikeholt.om Introdution Introdution This PDF is a free resoure from Mike Holt Enterprises, In. It s Unit 1 from the Eletrial

More information

Account Contract for Card Acceptance

Account Contract for Card Acceptance Aount Contrat for Card Aeptane This is an Aount Contrat for the aeptane of debit ards and redit ards via payment terminals, on the website and/or by telephone, mail or fax. You enter into this ontrat with

More information

computer science Program Educational Objectives

computer science Program Educational Objectives omputer siene bahelor of siene minor ertifiates: managing information on the world wide web master of siene in omputer siene master of siene in software engineering advaned ertifiate programs: bioinformatis

More information

INCOME TAX WITHHOLDING GUIDE FOR EMPLOYERS

INCOME TAX WITHHOLDING GUIDE FOR EMPLOYERS Virginia Department of Taxation INCOME TAX WITHHOLDING GUIDE FOR EMPLOYERS www.tax.virginia.gov 2614086 Rev. 01/16 Table of Contents Introdution... 1 Important... 1 Where to Get Assistane... 1 Online File

More information

A Three-Hybrid Treatment Method of the Compressor's Characteristic Line in Performance Prediction of Power Systems

A Three-Hybrid Treatment Method of the Compressor's Characteristic Line in Performance Prediction of Power Systems A Three-Hybrid Treatment Method of the Compressor's Charateristi Line in Performane Predition of Power Systems A Three-Hybrid Treatment Method of the Compressor's Charateristi Line in Performane Predition

More information

Context-Sensitive Adjustments of Cognitive Control: Conflict-Adaptation Effects Are Modulated by Processing Demands of the Ongoing Task

Context-Sensitive Adjustments of Cognitive Control: Conflict-Adaptation Effects Are Modulated by Processing Demands of the Ongoing Task Journal of Experimental Psyhology: Learning, Memory, and Cognition 2008, Vol. 34, No. 3, 712 718 Copyright 2008 by the Amerian Psyhologial Assoiation 0278-7393/08/$12.00 DOI: 10.1037/0278-7393.34.3.712

More information

Chapter 6 A N ovel Solution Of Linear Congruenes Proeedings NCUR IX. (1995), Vol. II, pp. 708{712 Jerey F. Gold Department of Mathematis, Department of Physis University of Utah Salt Lake City, Utah 84112

More information

Wireless Transmission Systems. Instructor

Wireless Transmission Systems. Instructor Teleommuniation Engineering Tehnology, Texas A&M University Wireless Transmission Systems Leture Notes - Copyright Jeff M. MDougall 2 Wireless Transmission Systems Teleommuniation Engineering Tehnology,

More information

The Optimal Deterrence of Tax Evasion: The Trade-off Between Information Reporting and Audits

The Optimal Deterrence of Tax Evasion: The Trade-off Between Information Reporting and Audits The Optimal Deterrene of Tax Evasion: The Trade-off Between Information Reporting and Audits Yulia Paramonova Department of Eonomis, University of Mihigan Otober 30, 2014 Abstrat Despite the widespread

More information

SCHEME FOR FINANCING SCHOOLS

SCHEME FOR FINANCING SCHOOLS SCHEME FOR FINANCING SCHOOLS UNDER SECTION 48 OF THE SCHOOL STANDARDS AND FRAMEWORK ACT 1998 DfE Approved - Marh 1999 With amendments Marh 2001, Marh 2002, April 2003, July 2004, Marh 2005, February 2007,

More information

Impact Simulation of Extreme Wind Generated Missiles on Radioactive Waste Storage Facilities

Impact Simulation of Extreme Wind Generated Missiles on Radioactive Waste Storage Facilities Impat Simulation of Extreme Wind Generated issiles on Radioative Waste Storage Failities G. Barbella Sogin S.p.A. Via Torino 6 00184 Rome (Italy), barbella@sogin.it Abstrat: The strutural design of temporary

More information

Findings and Recommendations

Findings and Recommendations Contrating Methods and Administration Findings and Reommendations Finding 9-1 ESD did not utilize a formal written pre-qualifiations proess for seleting experiened design onsultants. ESD hose onsultants

More information

Chapter 5 Single Phase Systems

Chapter 5 Single Phase Systems Chapter 5 Single Phase Systems Chemial engineering alulations rely heavily on the availability of physial properties of materials. There are three ommon methods used to find these properties. These inlude

More information

A Theoretical Analysis of Credit Card Reform in Australia *

A Theoretical Analysis of Credit Card Reform in Australia * A Theoretial Analysis of Credit Card Reform in Australia * by Joshua S. Gans and Stephen P. King Melbourne Business Shool University of Melbourne First Draft: 12 th May, 2001 This Version: 5 th May, 2003

More information

How To Fator

How To Fator CHAPTER hapter 4 > Make the Connetion 4 INTRODUCTION Developing seret odes is big business beause of the widespread use of omputers and the Internet. Corporations all over the world sell enryption systems

More information

FOOD FOR THOUGHT Topical Insights from our Subject Matter Experts

FOOD FOR THOUGHT Topical Insights from our Subject Matter Experts FOOD FOR THOUGHT Topial Insights from our Sujet Matter Experts DEGREE OF DIFFERENCE TESTING: AN ALTERNATIVE TO TRADITIONAL APPROACHES The NFL White Paper Series Volume 14, June 2014 Overview Differene

More information

UNIVERSITY AND WORK-STUDY EMPLOYERS WEB SITE USER S GUIDE

UNIVERSITY AND WORK-STUDY EMPLOYERS WEB SITE USER S GUIDE UNIVERSITY AND WORK-STUDY EMPLOYERS WEB SITE USER S GUIDE September 8, 2009 Table of Contents 1 Home 2 University 3 Your 4 Add 5 Managing 6 How 7 Viewing 8 Closing 9 Reposting Page 1 and Work-Study Employers

More information

Big Data Analysis and Reporting with Decision Tree Induction

Big Data Analysis and Reporting with Decision Tree Induction Big Data Analysis and Reporting with Deision Tree Indution PETRA PERNER Institute of Computer Vision and Applied Computer Sienes, IBaI Postbox 30 11 14, 04251 Leipzig GERMANY pperner@ibai-institut.de,

More information

A Survey of Usability Evaluation in Virtual Environments: Classi cation and Comparison of Methods

A Survey of Usability Evaluation in Virtual Environments: Classi cation and Comparison of Methods Doug A. Bowman bowman@vt.edu Department of Computer Siene Virginia Teh Joseph L. Gabbard Deborah Hix [ jgabbard, hix]@vt.edu Systems Researh Center Virginia Teh A Survey of Usability Evaluation in Virtual

More information

Procurement auctions are sometimes plagued with a chosen supplier s failing to accomplish a project successfully.

Procurement auctions are sometimes plagued with a chosen supplier s failing to accomplish a project successfully. Deision Analysis Vol. 7, No. 1, Marh 2010, pp. 23 39 issn 1545-8490 eissn 1545-8504 10 0701 0023 informs doi 10.1287/dea.1090.0155 2010 INFORMS Managing Projet Failure Risk Through Contingent Contrats

More information

Discovering Trends in Large Datasets Using Neural Networks

Discovering Trends in Large Datasets Using Neural Networks Disovering Trends in Large Datasets Using Neural Networks Khosrow Kaikhah, Ph.D. and Sandesh Doddameti Department of Computer Siene Texas State University San Maros, Texas 78666 Abstrat. A novel knowledge

More information

) ( )( ) ( ) ( )( ) ( ) ( ) (1)

) ( )( ) ( ) ( )( ) ( ) ( ) (1) OPEN CHANNEL FLOW Open hannel flow is haraterized by a surfae in ontat with a gas phase, allowing the fluid to take on shapes and undergo behavior that is impossible in a pipe or other filled onduit. Examples

More information

AngelCast: Cloud-based Peer-Assisted Live Streaming Using Optimized Multi-Tree Construction

AngelCast: Cloud-based Peer-Assisted Live Streaming Using Optimized Multi-Tree Construction AngelCast: Cloud-based Peer-Assisted Live Streaming Using Optimized Multi-Tree Constrution Raymond Sweha Boston University remos@s.bu.edu Vathe Ishakian Boston University visahak@s.bu.edu Azer Bestavros

More information

MEMBER. Application for election MEMBER, NEW GRADUATE. psychology.org.au. April 2015

MEMBER. Application for election MEMBER, NEW GRADUATE. psychology.org.au. April 2015 MEMBER Appliation for eletion MEMBER, NEW GRADUATE April 2015 psyhology.org.au MEMBER Belonging to the Australian Psyhologial Soiety (APS) means you are part of an ative, progressive organisation whih

More information

Availability, Reliability, Maintainability, and Capability

Availability, Reliability, Maintainability, and Capability Availability, Reliability, Maintainability, and Capability H. Paul Barringer, P.E. Barringer & Assoiates, In. Humble, TX Triplex Chapter Of The Vibrations Institute Hilton Hotel Beaumont, Texas February

More information

i e AT 8 of 1938 THE PERSONAL INJURIES (EMERGENCY PROVISIONS) ACT 1939

i e AT 8 of 1938 THE PERSONAL INJURIES (EMERGENCY PROVISIONS) ACT 1939 i e AT 8 of 1938 THE PERSONAL INJURIES (EMERGENCY PROVISIONS) ACT 1939 The Personal Injuries (Emergeny Provisions) At 1939 Index i e THE PERSONAL INJURIES (EMERGENCY PROVISIONS) ACT 1939 Index Setion

More information

THE UNIVERSITY OF TEXAS AT ARLINGTON COLLEGE OF NURSING. NURS 6390-004 Introduction to Genetics and Genomics SYLLABUS

THE UNIVERSITY OF TEXAS AT ARLINGTON COLLEGE OF NURSING. NURS 6390-004 Introduction to Genetics and Genomics SYLLABUS THE UNIVERSITY OF TEXAS AT ARLINGTON COLLEGE OF NURSING NURS 6390-004 Introdution to Genetis and Genomis SYLLABUS Summer Interession 2011 Classroom #: TBA and 119 (lab) The University of Texas at Arlington

More information

F220 Series. Installation Instructions. Photoelectric Smoke/Heat Detectors

F220 Series. Installation Instructions. Photoelectric Smoke/Heat Detectors F0 Series EN Installation Instrutions Photoeletri Smoke/Heat Detetors F0 Series Installation Instrutions.0 General Information EN.0 General Information. F0-B6 Series Bases Use with the F0 Series Heat and

More information

SUDOKU: Secure and Usable Deployment of Keys on Wireless Sensors

SUDOKU: Secure and Usable Deployment of Keys on Wireless Sensors SUDOKU: Seure and Usable Deployment of Keys on Wireless Sensors Matthias Wilhelm, Ivan Martinovi, Ersin Uzun, and Jens B. Shmitt diso Distributed Computer Systems Lab, TU Kaiserslautern, Germany {wilhelm,

More information

Capacity of Multi-Channel Wireless Networks. Random (c, f) Assignment

Capacity of Multi-Channel Wireless Networks. Random (c, f) Assignment Capaity of Multi-Channel Wireless Networks with Random (, f Assignment ABSTRACT Vartika Bhandari Dept. of Computer Siene, and Coordinated Siene Laboratory University of Illinois at Urbana-Champaign vbhandar@uiu.edu

More information