Capacity Planning of Survivable MPLS Networks Supporting DiffServ

Size: px
Start display at page:

Download "Capacity Planning of Survivable MPLS Networks Supporting DiffServ"

Transcription

1 Capacity Panning of Survivabe MPLS Networks Supporting DiffServ Kehang Wu and Dougas S. Reeves Departments of Eectrica and Computer Engineering and Computer Science North Caroina State University Abstract It is essentia for ISPs to offer both performance and survivabiity guarantees at the IP/MPLS ayer. Capacity panning is needed to ensure there wi be sufficient resource avaiabiity and quaity of service under norma circumstances, and during network faiures. In this paper we study the issue of capacity panning for survivabe MPLS networks providing DiffServ EF and BE traffic casses. Our goa is to minimize the tota ink cost, subject to the performance and survivabiity constraints of both EF and BE casses. The probem is formuated as an optimization probem, where we jointy seect the routes for edge to edge EF and BE user demand pairs, and assign a discrete capacity vaue for each ink. Capacity for the EF cass is fuy restorabe under the singe ink faiure mode, whie restorabiity of the BE cass can be adjusted by the network operator. We propose an efficient soution approach based on the Lagrangian Reaxation and subgradient methods. Computationa resuts show that the soution quaity is within a few percent of optima, whie the running time remains reasonabe for networks with 1000 nodes, 2500 inks, and 40,000 demand pairs. This represents the first work on capacity panning for survivabe mutipe-cass-of-service IP/MPLS networks with non-inear performance constraints. 1 Introduction The ast few years witnessed the exposive growth of mission-critica data traffic carried by the Internet, and the emergence of genera societa dependencies on the Internet. We are now amost as dependent on This work is supported by DARPA and AFOSR (under grants F and F ). 1

2 the avaiabiity of the communication networks as on other basic infrastructures, such as roads, water, and power. With up to 100 terabits per second of data traffic possiby carried in a singe fiber with DWDM, faiure of a singe fiber coud be catastrophic. Despite efforts to physicay protect the fiber optic cabes, cabe cuts are surprisingy frequent according to the FCC. Survivabe network design, which refers to the incorporation of survivabiity strategies to mitigate the impact of the given faiure scenarios, has become a vita part of network design, not an afterthought. The current network architecture is moving toward a two ayers structure where IP/MPLS [15] [29] is on top of optica WDM networks. Traditiony, network resiience primariy reies upon the functionaities provided by SONET and ATM [35][31]. As mentioned in [18], most ISPs rey on the IP ayer when faiures occur. But IP-based faiure recovery schemes suffer from ong deay and/or routing instabiity [27]. MPLS-based recovery has become a viabe soution because of the faster restoration time than IP rerouting. There are many works that investigated recovery schemes in MPLS enabed networks [14][20]. With MPLS ayer protection, the faiure is detected either by the MPLS ayer detection mechanism (such as exchange of Heo message ), or the signaing message propagated from the ower ayer [14]. Once a faiure is detected, protection switching based recovery mechanisms are used. A LSP is protected by a pre-estabished recovery path or path segment. The resource aocated to the recovery path, the socaed spare capacity, may be fuy avaiabe to preemptibe ow priority traffic. The spare capacity can be shared by mutipe recovery paths corresponding to different faiure scenarios, to further reduce the resource redundancy [19]. With MPLS-based recovery, it is possibe to differentiate the eve of protection for different casses of service [5]. For exampe, Expedited Forwarding (EF) based service, such as virtua ease ine service, can be supported by ink protection with a reserve path that offers 100% restorabiity. Best effort (BE) traffic, on the other hand, may not need to be fuy restorabe, or may simpy rey on IP rerouting without being assigned any spare capacity in the network. The cost of redundancy for survivabiity can be very high, compared to a corresponding network 2

3 designed ony to serve the working demands under norma conditions. One of the major interests in survivabe network design has been providing cost-efficient resource aocation. Most of the work on survivabe network design has been concentrated on the cassic spare capacity aocation probem (SCA) [13][23][22]. The SCA probem assumes that information about the operationa network is given; the focus is ony on the cacuation of backup paths and spare capacities, not the primary path and tota capacity. It has been shown in [19], however, that even though optimizing the working path and the backup path together is computationay more expensive, it offers significant advantage for the tota capacity savings. Since capacity panning is usuay performed infrequenty (on the time scae of weeks to months), we can afford the extra running time in exchange for a ower network cost. With the popuarization of e-commerce and new vaue-added services over IP, it is desirabe for the next generation Internet to offer Quaity of Service (QoS), the abiity to support mutipe traffic casses with different eves of performance guarantees. MPLS and Differentiated Service [8][28] are regarded as two key components of QoS. There has been work on survivabe network design touching on the issues of muticass traffic, such as the architectura aspects of differentiated survivabiity [5]. However, no quantitative treatment of the subject has been done so far. In this paper, we study the issue of capacity panning of survivabe MPLS networks. The targeted MPLS networks wi be providing DiffServ EF and BE traffic casses [34][11]. Since our approach requires a precise performance mode for optimization, we do not incude the AF traffic casses in this paper, due to the ack of a consensus on the impementation of the AF PHB [4]. We focus on the probem of resource dimensioning and traffic routing, and assume that the network topoogy is given. The probem is formuated as an optimization probem, where we jointy seect the primary and backup LSPs for edge to edge EF and BE user demand pairs 1, and assign a discrete capacity vaue for each ink. The goa is 1 The term demand pair and demand are used interchangeaby through out the paper. Depending on the goas of the network designer, a demand pair may correspond to a singe fow, or an aggregation of fows sharing the same QoS requirements, source and destination, and path through the network. 3

4 to minimize the tota ink cost, subject to the performance and survivabiity constraints of both EF and BE casses. The non-bifurcated (i.e., singe-path) routing mode is used for the EF cass as required by [11], so that the traffic from a singe EF demand pair wi foow the same LSP between the origin and the destination. Traffic in the BE cass is aowed to be spit across mutipe LSPs. The performance constraint of EF traffic is ony represented by a bandwidth requirement, as specified in [11]. Most of iterature on the design of survivabe network ony consider the case of singe ink faiures [14]. Whie there are recent works investigating the issues of doube-ink faiure [10], the existing methods for protection switching are designed for singe ink faiures. We assume there is ony singe ink faiure in this paper. We assume that each EF user demand is assigned a primary LSP as we as a backup LSP with enough preemptibe bandwidth; therefore it wi not be affected by any singe ink faiure. The performance as we as the survivabiity constraints of the BE cass are characterized by the upper imit on the average deay in each ink, assuming no more than a singe ink faiure. Queueing is modeed as M/G/1 strict priority queues. The nove aspect of our probem is the fact that two traffic casses, EF and BE, with independent performance and survivabiity requirements, share the same capacity resource, which resuts in a compex non-inear performance constraint. In addition to the non-inear performance constraint, non-bifurcated routing and discrete ink capacity constraints dramaticay increase the degree of difficuty, and significanty imit the viabe soution approaches. There are reated papers on capacity panning that do not dea with survivabiity issues. A survey of those works can be found in [32]. The remainder of this paper is organized as foows. In Section 2, notation and detaied assumptions and modes are presented. The probem definition is given in Section 3. Section 4 shows a Lagrangean reaxation of the origina probem, and describes the subgradient procedure to sove the resuting dua probem. Section 5 presents some numerica resuts on the use of the method. The paper is concuded in 4

5 Section 6. 2 Notation and Assumptions The issue of capacity panning of reiabe MPLS networks supporting DiffServ EF and BE traffic casses is studied in this paper. The network is considered reiabe as ong as the given performance constraints are satisfied under any singe ink faiure scenarios. A ink based formuation is used throughout this paper. Our notation is defined in Tabe 1. Compared to our previous work [34], B kj,δb, ˆxef jkb, ˆxbe jkb, and r are new variabes introduced to refect the survivabiity constraints. Pease refer to [32] for more compete discussion on the notations and assumptions. The network topoogy is assumed to be given, as we as the set of candidate LSPs J k for each O-D (Origin-Destination) pairs. As mentioned, we ony study the faiure scenario where there is at most a singe ink faiure. The given networks must be two-ink-connected [12], which impies that there are at east two ink-disjoint paths for any O-D pair. For every candidate path j, j J k, there is a set of set of candidate backup paths B kj, which consists of LSPs that are ink-disjoint from the primary LSP j, to ensure the avaiabiity of the backup path regardess of the faiure scenarios. δb defines the route of backup path b. The backup paths for the EF and BE user demands on path j, wi be specified by the backup path routing variabes ˆx ef jkb, ˆxbe jkb respectivey. T denotes the index of avaiabe ink types for ink. Ony one ink type can be used for each individua ink. The capacity and the cost of ink, ψ and C respectivey, are: ψ = t T u t ψ t, C = t T u t C t (1) There is no inear reationship assumed between C t and ψ t, therefore the ink cost is not necessariy a inear function of the ink capacity. For an EF demand m, m M k, we differentiate between the average arriva rate, α ef km, and the requested bandwidth, ρef km. ρef km is usuay a vaue between the average arriva 5

6 Network Candidate paths and routing L set of inks in the network J k set of possibe candidate LSP paths for the O-D pair k, k K K set of (both EF and BE) Origin- δj ink-path indicator; 1 if path j uses ink, Destination (O-D) pairs j J k,k K, L, 0 otherwise T index of avaiabe ink types for ink L B kj set of candidate backup paths, which are ink disjoint from j, j J k u t ink type decision variabe; 1 if ink type δb ink-path indicator; 1 if backup path b t is used for ink L, 0 other wise uses ink, b B kj,j J k,k K, L, ψ t size of the capacity of ink type t, t T x ef kmj 0 otherwise EF primary path routing variabe; 1 if EF demand m, m M k, k K uses path ψ tota capacity of ink, L j J k, 0 otherwise x be kj BE primary path routing variabe: the C t cost of the ink type t, t T, L portion of BE demand k that uses candidate path j, j J k. x be kj can be any rea vaue C tota cost of ink, L between 0 and 1 jkbbackup path routing variabe, 1 if backup path b B kj is used by path j J k EF demand reated BE demand reated M k set of EF demands for O-D pair k, k K αk be average arriva rate of a BE traffic demand, k K α ef km average arriva rate of an EF traffic de- β be average arriva rate of tota BE traffic demand on ink L mand m, m M k, k K d average deay experienced by BE traffic ρ ef km η ˆx ef jkb, ˆxbe d max on ink L maximum vaue of d aowed for ink L requested bandwidth of EF traffic demand m, m M k, k K tota requested bandwidth of EF demand r BE traffic restoration eve on ink L β ef average arriva rate of tota EF traffic de- g BE deay bound factor mand on ink L ỹ,ỹ2 the first and second moment of packet size, (units: bits & bits 2 ) Tabe 1: Notation rate and the peak rate. It is noted in [11] that the packets of the EF traffic cass beonging to the same fow shoud not be reordered. Consequenty, traffic from the same EF demand can not be separated into different LSPs. For the same reason, ony one backup path wi be used per EF user demand. Given a faiure of ink F, F L, the tota amount of EF user demand on ink is: η = k K ρ ef km (x ef m M k j J k kmj δ j + x ef kmj δf j ˆx ef jkb δ b ) (2) where (x ef kmj δ j ) equas 1 if the EF demand m uses LSP j as its working path and the path j incudes ink, 6

7 and 0 otherwise, whie (x ef kmj δf j ˆxef jkb δ b ) equas 1 ony if the working path of the EF demand m is interrupted by the ink faiure F and the backup path b uses ink. Simiary, average arriva rate of tota EF traffic on ink can be cacuated according to the foowing equation: β ef = α ef km (x ef kmj δ j + x ef kmj δf j ˆx ef jkb δ b ) (3) k K m M k j J k For each O-D pair k, ony one BE demand pair is defined, which can be ooked as the aggregation of mutipe BE demands with the same O-D. Unike EF traffic, we aow the traffic within a singe BE demand to spit arbitrariy across any number of candidate LSPs, therefore the aggregation of BE traffic woud potentiay improve the effectiveness of traffic engineering. However, we assume ony one backup path wi be used for traffic on the same working path. We assume that the BE traffic may not be fuy restorabe in case of ink faiure. BE traffic restoration eve r, a vaue between [0,1], represents the proportion of BE traffic on each ink that is being protected. The abiity to set the restoration eve of BE traffic and accordingy differentiate the resiience eve provides a powerfu too to the service provider. It aows the ISP to ceary demonstrate the advantage of premium service based on EF cass, and aso aows the ISP to fine tune the redundancy eve and tota network cost. The tota BE oad (average arriva rate) on ink is: β be = k K α be k j J k (x be kj δ j + r x be kj δ F j ˆx be jkb δ b ), F L (4) It is assumed that the requested bandwidth of a EF user demand is defined in a way such that the performance of an EF demand wi be guaranteed if it is given the requested bandwidth. η ψ, L (5) There are many discussions regarding the imits on the ink utiization of EF user demands [9]. The revised EF PHB, RFC3246 [11], introduces an error term E a for the treatment of the EF aggregate, which represents the aowed worst case deviation between the actua EF packet departure time and the idea 7

8 departure time of the same packet. This revision makes it possibe for the EF utiization to go up as high as 100% [37]. In this paper, we assume that there are no additiona constraints on the EF utiization. We pick the average deay as the soe performance measurement for BE traffic in this paper as suggested by [24]. We evauate the performance of BE traffic on a per-ink basis (i.e., not end-to-end). The vaue ỹ ψ stands for the average transmission deay of packets. We use ỹ ψ as the basis for the deay bound. Let d max = g ỹ ψ, where g is a parameter defined by the network designer. The arger the vaue of g, the more bandwidth is required for ink, therefore the ower the ink utiization. We assume that the performance of BE traffic is satisfactory if d d max. Note that the deay bound d max remains the same under the norma condition and the faiure scenarios. Every router is modeed as a M/G/1 system with Poisson packet arrivas and an arbitrary packet ength distribution. [30] concudes, through both simuation and anaytic study, that even though the traffic exhibits bursty behavior at certain time scaes, the reationship between the variance and the mean is approximatey inear, or Poisson-ike, if they are measured over arger time scaes, where the traffic can be treated as if it were smooth. Our choice of the Poisson arriva mode is justified because we are more concerned about the average BE performance over a arge time scae for capacity panning purposes. From the average queueing deay formua of the priority queue [21], we obtain the performance constraint for BE traffic: d = ỹ ỹ + 2 ψ 2ỹ With some rearrangement, (6) yieds ( ψ β ef β ef + β be )( ψ β ef β be ) g ỹ ψ (6) ψ f(β ef,β be ) (7) where f(β ef,β be ) = β ef + βbe 2 + ỹ2 (β ef + β be ) 4(ỹ) 2 (g 1) (2β ef + β be + ỹ 2 (β ef + β be ) 2(ỹ) 2 (g 1) )2 4β be (β be + β ef ). 8

9 In order to have a meaningfu soution for constraint (7), ψ >β ef + β be is required. 3 Probem Formuation The forma probem definition is presented beow. Given: L, K, T,ψ t,c t,j k,δ j,b kj,δ b,m k,α ef km,ρef km,αbe k,d max,r,g, ỹ, ỹ2 Variabe: u t,x ef kmj,xbe kj, ˆxef jbk, ˆxbe jbk Goa: min L C t Subject to: u t =0/1, t T u t = 1 (8) x ef kmj =0/1, x ef kmj = 1 (9) j J k x be kj = 1 (10) j J k ˆx ef jkb =0/1, ˆx ef jkb = 1 (11) ˆx be jkb =0/1, ˆx be jkb = 1 (12) ψ η (13) ψ f(β ef,β be ) (14) (8) imposes a discrete constraint on the ink capacities. Constraint (9) ensures that a traffic from one EF O-D pair wi foow one singe path. (11)(12) denotes that ony one backup path wi be used for each working path. Constraint (13)(14) ensures the performance of EF and BE traffic respectivey. Because C is non-decreasing with respect to β ef, η, and β be, this probem can be reformuated as: min L C (15) 9

10 Subject to (8)(9)(10)(11)(12) (13)(14) and: β ef η k K k K β be k K ρ ef km (x ef m M k j J k α ef km (x ef m M k j J k α be k j J k (x be kj δ j + r kmj δ j + kmj δ j + x be x ef kmj δf j ˆx ef x ef kmj δf j ˆx ef jkb δ b ) (16) jkb δ b ) (17) kj δj F ˆx be jkb δ b ) (18) We refer to the probem defined by (15,8,9,10,11,12,13,14,16, 17, 18) as probem (P) in the rest of this paper. As can be seen from the above probem formuation, probem (P) is a non-inear integer programming probem, which is considered to be at east NP hard [2]. 4 Soution Method 4.1 Lagrangean Reaxation Lagrangean Reaxation is a common technique for muticommodity fow probems [2]. It has been successfuy appied to the capacity panning and routing probems [16] [17] [3] [25] [26], and our previous work [33][34]. We describe its use for our probem in this section. Pease refer to [32] for more detai. Using Lagrangean Reaxation, reax (16)(17) and (18), and we have the Lagrangean as: = L(x ef kmj,xbe kj, ˆxef jbk, ˆxbe jbk,u t,λ ef ( C λ η η λ ef β ef λ be β be ) + k K + k K m M k j J k x be x ef kmj j J k kj L The Lagrangean dua probem (D) is then:,λ η,λbe ) (λ η ρef kj + λef L α ef km )(δ j + δ F j ˆx ef jkb δ b ) λ be α ef k (δ j + r δ F j ˆx be jkb δ b ) (19) λ η,λef max,λ be 0 h(λ η,λef,λ be ) (20) 10

11 where: h(λ η,λef,λ be ) = min x ef kmj,xbe kj,ˆxef jbk,ˆxbe jbk,u t L(x ef kmj,xbe kj, ˆxef jbk, ˆxbe Since η,β ef,β be and x ef kmj,xbe kj, ˆxef jkb, ˆxbe jkb are independent variabes, min L = min( C λ η η λ ef β ef λ be β be ) + min[ x ef kmj (λ η ρef k K m M k j J k L + min[ x be kj k K j J k L 4.2 Soving the Subprobems kj + λef jbk,u t,λ ef,λ η,λbe ) (21) α ef km )(δ j + δ F j ˆx ef jkb δ b )] λ be α ef k (δ j + r δj F ˆx be jkb δb )] (22) Equation (22) shows that the probem (21) can be separated into the foowing three set of subprobems: Subprobem (i): min( C λ η η λ ef β ef λ be β be ) (23) Subject to: (8)(13)(14) (24) Subprobem (i) can be soved by the gradient projection method [7]. Subprobem (ii): min[ j J k x ef kmj (λ η ρef kj + λef L α ef km )(δ j + δj F ˆx ef jkb δ b )] (25) Since we have a sma set of candidate path for each O-D, we can simpy set the ink cost to (λ η ρef kj + λ ef α ef km ), and then enumerate the choice of working paths and backup paths, picking the combination of working and backup paths with the east tota cost. The resuting choice of working path and backup path is cached. In the subsequent iteration, ony paths whose cost is changed wi be compared to the previous soution, and hence save the running time. Subprobem (iii): min[ j J k x be kj L λ be α ef k (δ j + r δj F ˆx be jkb δb )] (26) 11

12 Simiar to Subprobem (ii), the soution is to enumerate the choice of working path and backup path, then choose the one with the east combined cost. The resuts are simiary cached to speed up the program. 4.3 Subgradient Method The subgradient method is used to update λ η, λef and λ be. For an in depth description of the subgradient method and its variants, pease refer to [6]. Given the initia vaues of λ η, λef and λ be, once we sove the probem (D), subgradients for the three set of Lagrangeans are computed as foows: ω η = η k K ω ef = β ef k K ω be = β be k K ρ ef km (x ef m M k j J k α ef km (x ef m M k j J k α be k j J k (x be kmj δ j + kj δ j + r x ef kmj δf j ˆx ef kmj δ j + x be x ef kmj δf j ˆx ef jkb δ b ) (27) jkb δ b ) (28) kj δj F ˆx be jkb δ b ) (29) To improve the convergence speed, the surrogate subgradient method is used, where ony one of the three subprobems is soved at a time. The subsequent Lagrangean mutipiers are updated using the atest vaue. λ η min(0,λ η + tωη ), L (30) λ ef min(0,λ ef + tω ef ), L (31) λ be min(0,λ be + tω be ), L (32) where the step size, t, is defined by: t = φ ω eta h h 2 + ω ef (33) 2 + ω be 2 where h is the vaue of the best feasibe soution found so far, and φ is a scaar between 0 and 2. φ is set to 2 initiay in our study and is haved if the soution does not improve in 10 iterations. 12

13 At each iteration, the soution of x ef kmj,x be kj, ˆx ef jkb, and ˆxef jkb for the prima probem (P) can be obtained from the soution of subprobem (ii) and(iii). The ink capacity ψ is computed according to (13). Consequenty, the primary objective function can be derived. As the iteration proceeds, we store the best soution found so far for the prima probem (P). In this way, we are aways abe to obtain a feasibe soution. The maximum number of iterations is set to 400 in the impementation [33][2]. The soution of the dua probem provides a ower bound for the prima probem. Therefore, the soution quaity can be assessed by the duaity gap, which is the difference between the soutions of probem (P) and probem (D). Note that because the duaity gap is aways no smaer than the actua difference between the obtained feasibe soution and the optima soution, it is a conservative estimate of the soution quaity. 5 Computationa Resuts In this section, we present numerica resuts based on experimentation. The objective of our experiment is to evauate the soution quaity and running time of the agorithm. The program is impemented in C and the computation is performed on a Pentium IV 2.4GHz PC with 512M memory, running Redhat Linux 7.2. The network topoogies are generated using the Georgia Tech Internetwork Topoogy Modes (GT-ITM) [38]. The ocations of origins and destinations are randomy seected. For each O-D pair, 10 candidate working paths are cacuated using Yen s K-shortest path agorithm [36]. 10 candidate backup paths are aso generated for each candidate working path. If not specified, EF demand pairs are randomy generated with the average rate uniformy distributed from 0 Mbps to 10 Mbps. The requested bandwidth of a EF demand pair is a vaue between the average rate and the peak rate. It is randomy specified to be 150%-300% of the average rate in our experiments. The average rate of BE demand is uniformy distributed from 10Mbps to 50Mbps. The number of EF demands for the same O-D pair is uniformy distributed from 1 to 10. The number of candidate ink types 13

14 for ink is uniformy distributed from 5 to 10. The capacities of ink types are set to be mutipes of 45Mbps, whie the costs of the ink types for ink are randomy generated in such a way that the cost goes higher and the unit cost per Mbps goes down as the ink capacity increases. We use average packet size ỹ = 4396 bits and second moment of packet size ỹ2 = bits 2 for a the test cases. They are cacuated based on a traffic trace (AIX ) from the NLANR Passive Measurement and Anaysis project [1]. The restoration eve of BE traffic is determined by the network designer, depending on the budget and the desired BE cass resiience. It is set to 0.5 in a of the experiments. The BE deay bound factor, g, is set to 2 for a inks in our experiments. In practice, g shoud be carefuy chosen to refect the desired BE performance and the expected ink utiization. Figure (1) shows the figure of f(β ef ) with respect to β ef for various vaues of g. As can be seen from the figure, the ink utiization varies significanty as g changes. The average ink utiization, the vaue of β ef +β be, is about f(βef ) 60% when g equas x 107 bps g = f ( β ef ) 10 Utiization = g = 2 g = 4 g = x 10 6 bps β ef Figure 1: f(β ef ) vs. β ef, (β be =5 107 bps) Two additiona greedy type heuristics were impemented for the purpose of comparison: 14

15 1. The first heuristic method starts with one randomy chosen O-D pair k, finds out both the best working path j and backup path b for both EF and BE demand among a candidate working and backup paths that woud resut in the east tota cost, and then repeats the process for a k K. 2. The second heuristic is an iterative method inspired by the heuristic used in [23]. The primary difference here is that [23] deas ony with the spare capacity panning probem, whie we have to find out the working path first. In the first iteration, for each O-D pair k, the agorithm finds out ony the best working paths j for both EF and BE demands among candidate working paths that woud resut in the east tota cost. After a the working paths have been determined, the agorithm ooks for the best backup paths b for both EF and BE demands among candidate backup paths. In the subsequent iteration, the agorithm goes through each O-D pair to see whether any change of working path or backup path wi ower the tota cost. The iteration wi continue as ong as there is at east one adjustment of path in the current iteration. We wi ca these two methods (H) and (S), respectivey. The agorithms were tested on 8 different sizes of networks, ranging from 10 nodes to 1000 nodes. Some detais of the network topoogies are isted in Tabe 2. Note that the O-D demand number shown in Tabe 2 incudes both EF and BE demands. To obtain confidence intervas, we generated 30 different topoogies for each network size, with the same number of nodes, inks, and O-D pairs. Duaity Gap = s p s d s d (34) The soution quaity is represented by the duaity gap, which is the percentage difference between the soution of the prima probem and the dua probem. s p and s d are the soutions of prima probem and dua probem respectivey. A vaue cose to zero means the soution is very cose to optima. Tabe 2 shows the Duaity Gaps of the three soution methods, expressed in terms of the 95% confidence intervas. It is easy to see from the tabe that the Lagrangian Reaxation together with the subgradient 15

16 Node Link O-D Duaity (%) Number Number Number LR H S Tabe 2: Network topoogy information and experimenta resuts method produces reasonabe resuts as the duaity gap is bounded by no more than 3.3%. Note the prima probem itsef is approximated when reducing the size of candidate path set for a possibe path set. But according to our experimenta resuts, more than 97% of the time, the fina soution is chosen among the 5 shortest candidate paths. Therefore, 10 candidate paths are considered adequate. Having more than 10 candidate paths wi have minima impact on the soution quaity, whie significanty increasing the running time. Given the arge number of networks being tested, we are confident that the soution shoud have good quaity for other sizes of networks. Heuristic (S) achieves comparabe soution quaity in networks sized up to 200 nodes, whie method (H) yieds soutions with a duaity gap of more than 20%, which is not very desirabe. Figure 2 shows the running time with respect to the network size. In a 240 test cases of Lagrangian Reaxation based method, the agorithm converges without difficuty. As can be seen from the figure, heuristic (S) takes much onger time than the other two methods. From the experiments shown above, Lagrangian Reaxation based approach demonstrates a good tradeoff between the soution quaity and the running time. The size of the argest network evauated in this paper is representative of a arge network, and is much arger than the test cases used in most work on network design. It is fair to predict that the running time of the agorithm wi stay reasonabe for practica sized networks. 16

17 9 x H LR S 9 x H LR S 7 7 Running time (sec) Running time (sec) Node number O-D pair number Figure 2: Running Time vs. Network Size The fow aggregation based agrangian Reaxation method proposed in this paper can be appied to more genera noninear integer muticommodity fow probems. We are seeking the opptunitues to extend the method to other types of network design probems, such as design of optica networks and overay networks. 6 Concusions and Future Directions In this paper, we addressed the probem of ink dimensioning and routing for survivabe MPLS networks supporting DiffServ EF and BE traffic. We formuate the probem as an optimization probem, where the tota ink cost is minimized, subject to the performance constraints of both EF and BE casses under norma circumstance and singe ink faiures. The variabe here is the routing of working and backup LSP of both EF and BE user demands, and the discrete ink capacities. This paper presents a preiminary investigation of the capacity panning issue for survivabe muticass MPLS networks. The novety of the probem presented in this paper is that it invoves two traffic casses, 17

18 EF and BE, which have totay different forms of performance requirements and resuts in non-inear performance constraint. Unike SCA probems, this paper dea with the the working and backup paths at the same time, which is considered to be much more compex. The combination of targeting mutipe traffic casses with non-inear performance constraints, optimizing working path and backup path simutaneousy, and the use of non-inear cost function produces a probem more difficut than most reated works. We presented a Lagrangian Reaxation-based method to effectivey decompose the origina probem. A subgradient method is used to find the optima Lagrangian mutipier. We investigated experimentay the soution quaity and running time of this approach. The resuts from our experiments indicate that our method produces soutions that are within a few percent of the optima soution, whie the running time stays reasonabe for practica sized networks. The probem formuation and soution approaches may be appied to other traffic casses, such as the Assured Forwarding (AF) cass. The technique of reaxing fow baance equations enabes the Lagrangian Reaxation techniques to be appied to difficut integer non-inear programming probems, which is not possibe otherwise. We are investigating the adaptation of this technique to the genera non-inear muticommodity fow probems. References [1] NLANR passive measurement and anaysis project, [2] R.K. Ahuja, T.L. Magnanti, and J.B. Orin. Network Fows, Theory, Agorithms and Appications. Prentice Ha, New Jersey, [3] A. Amiri. A System for the Design of Packet-Switched Communication Networks with Economic Tradeoffs. Computer Communications, 21(18): , Dec [4] Shigehiro Ano, Toru Hasegawa, and Nicoas Decre. Experimenta TCP performance evauation on diffserv AF PHBs over ATM SBR service. Teecommunication Systems, 19(3-4): , Apr [5] Achim Authenrieth and Andreas Kirstadter. Components of MPLS recovery supporting differentiated resiience requirements. In IFIP Workshop on IP and ATM Traffic Management WATM 2001, Paris, France, September [6] D.P. Bertsekas. Noninear Programming. Athena Scientific, Bemont, MA, [7] D.P. Bertsekas and R.G. Gaager. Data Networks. Prentice Ha, Engewood Ciffs, NJ, [8] S. Bake and et a. An Architecture for Differentiated Service, IETF RFC Dec

19 [9] A. Charny and J.-Y. Le Boudec. Deay Bounds in Network with Aggregate Scheduing. In First Internationa Workshop on Quaity of Future Internet Service, pages 1 13, Berin, Germany, Sep [10] Hongsik Choi, Suresh Subramaniam, and Hyeong-Ah Choi. On doube-ink faiure recovery in wdm optica networks. In Proceeding of IEEE Infocom 02, June [11] B. David, A. Charny, and et A. An Expedited Forwarding PHB, IETF RFC Mar [12] Reinhard Dieste. Graph Theory. Graduate Textbooks in Mathematics 173. Springer-Verag, 2 edition, [13] C. Dovrois and P. Ramanathan. Resource aggregation for faut toerance in integrated service networks. ACM Computer Communication Review, 28(2):39 53, [14] C. Huang et A. Buiding reiabe MPLS networks using a path protection mechanism. IEEE Communications Magazine, pages , Mar [15] D. Awduche et a. Requirement for traffic engineering over MPLS, IETF RFC September [16] B. Gavish and I. Neuman. Capacity and Fow Assignment in Large Computer Networks. In IEEE Infocom 86, pages IEEE, Apr [17] B. Gavish and I. Neuman. A System for Routing and Capacity Assignment in Computer Networks. IEEE Transactions on Communications, 37(4): , Apr [18] F. Giroire, A. Nucci, N. Taft, and C. Diot. Increasing the robustness of IP backbones in the absence of optica eve protection. In Infocom 03. IEEE, May [19] R. Iraschko, M. MacGregor, and W. Grover. Optima capacity pacement for path resotration in STM or ATM mesh survivabe networks. IEEE/ACM transcations on Networking, 6(3): , June [20] S. Kim and K. G. Shin. Improving dependabiity of rea-time communication with prepanned backup routes and spare resource poo. In Proceeding of IWQoS 2003, June [21] L. Keinrock. Queueing Systems, Voume II: Computer Appication. Wiey Interscience, [22] M. Kodiaam and T.V. Lakshman. Dynamic routing of bandwidth guaranteed tunnes with resotration. In Proceeding of IEEE Infocom 00, Mar [23] Y. Liu and D. Tipper. Approximation optima spare capacity aocation by successive survivabe routing. In Proceeding of IEEE Infocom 01, Anchorage, AL, Apri [24] Jim Martin and Arne Nisson. On service eve agreements for IP networks. In IEEE Infocom 02, New York, IEEE. [25] D. Medhi. Muti-Hour, Muti-Traffic Cass Network Design for Virtua Path-based Dynamicay Reconfigurabe Wide-Area ATM Networks. IEEE/ACM Transactions on Networking, 3(6): , Dec [26] D. Medhi and D. Tipper. Some Approaches to Soving a Mutihour Broadband Network Capacity Design Probem with Singe-path Routing. Teecommunication Systems, 13(2-4): , [27] S. Neakuditi, S. Lee, Y. Yu, and Zhi-Li Zhang. Faiure insensitive routing for ensuring service avaiabiity. In Proceeding of IWQoS 2003, June [28] K. Nichos, V. Jacobson, and L. Zhang. A Two-bit Differentiated Services Architecture for the Internet, IETF RFC Ju [29] E. Rosen, A. Viswanathan, and R. Caon. Mutiprotoco abe switching architecture, IETF RFC January [30] Xusheng Tian, Jie Wu, and Chuangyi Ji. A unified framework for understanding network traffic using independent waveet modes. In IEEE Infocom 02, voume 1, pages , New York, IEEE. [31] Pau Veitech and Dave Johnson. ATM network resiience. pages 26 23, September/October [32] Kehang Wu. Fow Aggregation based Lagrangian Reaxation with Appications to Capacity Panning of IP Networks with Mutipe Casses of Service. PhD thesis, North Caroina State University, Raeigh,

20 [33] Kehang Wu and Dougas S. Reeves. Capacity panning of DiffServ networks with Best-Effort and Expedited Forwarding traffic. In IEEE 2003 Internationa Conference on Communications, Anchorage, Aaska, May [34] Kehang Wu and Dougas S. Reeves. Link Dimensioning and LSP Optimization for MPLS Networks Supporting DiffServ EF and BE traffic casses, [35] Tsong-Ho Wu. Fiber Network Service Survivabiity. Artech House, [36] J.Y. Yen. Finding the K Shortest Loopess Paths in a Network. Management Science, 17(11): , Ju [37] Jiang Yuming. Deay bounds for a network of guaranteed rate servers with FIFO aggregation. Computer Networks, 40(6): , December [38] E.W. Zegura, K. Cavert, and M. Jeff Donahoo. A Quantitative Comparison of Graph-based Modes for Internet Topoogy. IEEE/ACM Transactions on Networks, 5(6): , Dec

Link Dimensioning and LSP Optimization for MPLS Networks Supporting DiffServ EF and BE traffic classes

Link Dimensioning and LSP Optimization for MPLS Networks Supporting DiffServ EF and BE traffic classes Link Dimensioning and LSP Optimization for MPLS Networks Supporting DiffServ EF and BE traffic casses Kehang Wu and Dougas S. Reeves Departments of Eectrica and Computer Engineering and Computer Science

More information

Pricing Internet Services With Multiple Providers

Pricing Internet Services With Multiple Providers Pricing Internet Services With Mutipe Providers Linhai He and Jean Warand Dept. of Eectrica Engineering and Computer Science University of Caifornia at Berkeey Berkeey, CA 94709 inhai, wr@eecs.berkeey.edu

More information

Simultaneous Routing and Power Allocation in CDMA Wireless Data Networks

Simultaneous Routing and Power Allocation in CDMA Wireless Data Networks Simutaneous Routing and Power Aocation in CDMA Wireess Data Networks Mikae Johansson *,LinXiao and Stephen Boyd * Department of Signas, Sensors and Systems Roya Institute of Technoogy, SE 00 Stockhom,

More information

Secure Network Coding with a Cost Criterion

Secure Network Coding with a Cost Criterion Secure Network Coding with a Cost Criterion Jianong Tan, Murie Médard Laboratory for Information and Decision Systems Massachusetts Institute of Technoogy Cambridge, MA 0239, USA E-mai: {jianong, medard}@mit.edu

More information

IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 31, NO. 12, DECEMBER 2013 1

IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 31, NO. 12, DECEMBER 2013 1 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 31, NO. 12, DECEMBER 2013 1 Scaabe Muti-Cass Traffic Management in Data Center Backbone Networks Amitabha Ghosh, Sangtae Ha, Edward Crabbe, and Jennifer

More information

GREEN: An Active Queue Management Algorithm for a Self Managed Internet

GREEN: An Active Queue Management Algorithm for a Self Managed Internet : An Active Queue Management Agorithm for a Sef Managed Internet Bartek Wydrowski and Moshe Zukerman ARC Specia Research Centre for Utra-Broadband Information Networks, EEE Department, The University of

More information

Lecture 7 Datalink Ethernet, Home. Datalink Layer Architectures

Lecture 7 Datalink Ethernet, Home. Datalink Layer Architectures Lecture 7 Dataink Ethernet, Home Peter Steenkiste Schoo of Computer Science Department of Eectrica and Computer Engineering Carnegie Meon University 15-441 Networking, Spring 2004 http://www.cs.cmu.edu/~prs/15-441

More information

Multi-Robot Task Scheduling

Multi-Robot Task Scheduling Proc of IEEE Internationa Conference on Robotics and Automation, Karsruhe, Germany, 013 Muti-Robot Tas Scheduing Yu Zhang and Lynne E Parer Abstract The scheduing probem has been studied extensivey in

More information

Pricing and Revenue Sharing Strategies for Internet Service Providers

Pricing and Revenue Sharing Strategies for Internet Service Providers Pricing and Revenue Sharing Strategies for Internet Service Providers Linhai He and Jean Warand Department of Eectrica Engineering and Computer Sciences University of Caifornia at Berkeey {inhai,wr}@eecs.berkeey.edu

More information

An Idiot s guide to Support vector machines (SVMs)

An Idiot s guide to Support vector machines (SVMs) An Idiot s guide to Support vector machines (SVMs) R. Berwick, Viage Idiot SVMs: A New Generation of Learning Agorithms Pre 1980: Amost a earning methods earned inear decision surfaces. Linear earning

More information

Teamwork. Abstract. 2.1 Overview

Teamwork. Abstract. 2.1 Overview 2 Teamwork Abstract This chapter presents one of the basic eements of software projects teamwork. It addresses how to buid teams in a way that promotes team members accountabiity and responsibiity, and

More information

Virtual trunk simulation

Virtual trunk simulation Virtua trunk simuation Samui Aato * Laboratory of Teecommunications Technoogy Hesinki University of Technoogy Sivia Giordano Laboratoire de Reseaux de Communication Ecoe Poytechnique Federae de Lausanne

More information

Cooperative Content Distribution and Traffic Engineering in an ISP Network

Cooperative Content Distribution and Traffic Engineering in an ISP Network Cooperative Content Distribution and Traffic Engineering in an ISP Network Wenjie Jiang, Rui Zhang-Shen, Jennifer Rexford, Mung Chiang Department of Computer Science, and Department of Eectrica Engineering

More information

Online Supplement for The Robust Network Loading Problem under Hose Demand Uncertainty: Formulation, Polyhedral Analysis, and Computations

Online Supplement for The Robust Network Loading Problem under Hose Demand Uncertainty: Formulation, Polyhedral Analysis, and Computations Onine Suppement for The Robust Network Loading Probem under Hose Demand Uncertaint: Formuation, Pohedra Anasis, and Computations Aşegü Atın Department of Industria Engineering, TOBB Universit of Economics

More information

Maintenance activities planning and grouping for complex structure systems

Maintenance activities planning and grouping for complex structure systems Maintenance activities panning and grouping for compex structure systems Hai Canh u, Phuc Do an, Anne Barros, Christophe Berenguer To cite this version: Hai Canh u, Phuc Do an, Anne Barros, Christophe

More information

Fast Robust Hashing. ) [7] will be re-mapped (and therefore discarded), due to the load-balancing property of hashing.

Fast Robust Hashing. ) [7] will be re-mapped (and therefore discarded), due to the load-balancing property of hashing. Fast Robust Hashing Manue Urueña, David Larrabeiti and Pabo Serrano Universidad Caros III de Madrid E-89 Leganés (Madrid), Spain Emai: {muruenya,darra,pabo}@it.uc3m.es Abstract As statefu fow-aware services

More information

TOPOLOGICAL DESIGN OF MULTIPLE VPNS OVER MPLS NETWORK Anotai Srikitja David Tipper

TOPOLOGICAL DESIGN OF MULTIPLE VPNS OVER MPLS NETWORK Anotai Srikitja David Tipper TOPOLOGICAL DESIGN OF MULTIPLE VPNS OVER MPLS NETWORK Anotai Sriitja David Tier Det. of Information Science and Teecommunications University of Pittsburgh N. Beefied Avenue, Pittsburgh, PA 60 ABSTRACT

More information

GreenTE: Power-Aware Traffic Engineering

GreenTE: Power-Aware Traffic Engineering GreenTE: Power-Aware Traffic Engineering Mingui Zhang zmg6@mais.tsinghua.edu.cn Cheng Yi yic@emai.arizona.edu Bin Liu iub@tsinghua.edu.cn Beichuan Zhang bzhang@arizona.edu Abstract Current network infrastructures

More information

WIRELESS Mesh Networks (WMNs) have recently attracted

WIRELESS Mesh Networks (WMNs) have recently attracted 3968 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 2, NO. 8, AUGUST 23 A New MPLS-Based Forwarding Paradigm for Muti-Radio Wireess Mesh Networks Stefano Avaone and Giovanni Di Stasi Abstract Routing

More information

Load Balancing in Distributed Web Server Systems with Partial Document Replication *

Load Balancing in Distributed Web Server Systems with Partial Document Replication * Load Baancing in Distributed Web Server Systems with Partia Document Repication * Ling Zhuo Cho-Li Wang Francis C. M. Lau Department of Computer Science and Information Systems The University of Hong Kong

More information

Comparison of Traditional and Open-Access Appointment Scheduling for Exponentially Distributed Service Time

Comparison of Traditional and Open-Access Appointment Scheduling for Exponentially Distributed Service Time Journa of Heathcare Engineering Vo. 6 No. 3 Page 34 376 34 Comparison of Traditiona and Open-Access Appointment Scheduing for Exponentiay Distributed Service Chongjun Yan, PhD; Jiafu Tang *, PhD; Bowen

More information

Australian Bureau of Statistics Management of Business Providers

Australian Bureau of Statistics Management of Business Providers Purpose Austraian Bureau of Statistics Management of Business Providers 1 The principa objective of the Austraian Bureau of Statistics (ABS) in respect of business providers is to impose the owest oad

More information

A Supplier Evaluation System for Automotive Industry According To Iso/Ts 16949 Requirements

A Supplier Evaluation System for Automotive Industry According To Iso/Ts 16949 Requirements A Suppier Evauation System for Automotive Industry According To Iso/Ts 16949 Requirements DILEK PINAR ÖZTOP 1, ASLI AKSOY 2,*, NURSEL ÖZTÜRK 2 1 HONDA TR Purchasing Department, 41480, Çayırova - Gebze,

More information

The Use of Cooling-Factor Curves for Coordinating Fuses and Reclosers

The Use of Cooling-Factor Curves for Coordinating Fuses and Reclosers he Use of ooing-factor urves for oordinating Fuses and Recosers arey J. ook Senior Member, IEEE S& Eectric ompany hicago, Iinois bstract his paper describes how to precisey coordinate distribution feeder

More information

Face Hallucination and Recognition

Face Hallucination and Recognition Face Haucination and Recognition Xiaogang Wang and Xiaoou Tang Department of Information Engineering, The Chinese University of Hong Kong {xgwang1, xtang}@ie.cuhk.edu.hk http://mmab.ie.cuhk.edu.hk Abstract.

More information

WINMAG Graphics Management System

WINMAG Graphics Management System SECTION 10: page 1 Section 10: by Honeywe WINMAG Graphics Management System Contents What is WINMAG? WINMAG Text and Graphics WINMAG Text Ony Scenarios Fire/Emergency Management of Fauts & Disabement Historic

More information

Minimizing the Total Weighted Completion Time of Coflows in Datacenter Networks

Minimizing the Total Weighted Completion Time of Coflows in Datacenter Networks Minimizing the Tota Weighted Competion Time of Cofows in Datacenter Networks Zhen Qiu Ciff Stein and Yuan Zhong ABSTRACT Communications in datacenter jobs (such as the shuffe operations in MapReduce appications

More information

TERM INSURANCE CALCULATION ILLUSTRATED. This is the U.S. Social Security Life Table, based on year 2007.

TERM INSURANCE CALCULATION ILLUSTRATED. This is the U.S. Social Security Life Table, based on year 2007. This is the U.S. Socia Security Life Tabe, based on year 2007. This is avaiabe at http://www.ssa.gov/oact/stats/tabe4c6.htm. The ife eperiences of maes and femaes are different, and we usuay do separate

More information

Scheduling in Multi-Channel Wireless Networks

Scheduling in Multi-Channel Wireless Networks Scheduing in Muti-Channe Wireess Networks Vartika Bhandari and Nitin H. Vaidya University of Iinois at Urbana-Champaign, USA vartikab@acm.org, nhv@iinois.edu Abstract. The avaiabiity of mutipe orthogona

More information

Normalization of Database Tables. Functional Dependency. Examples of Functional Dependencies: So Now what is Normalization? Transitive Dependencies

Normalization of Database Tables. Functional Dependency. Examples of Functional Dependencies: So Now what is Normalization? Transitive Dependencies ISM 602 Dr. Hamid Nemati Objectives The idea Dependencies Attributes and Design Understand concepts normaization (Higher-Leve Norma Forms) Learn how to normaize tabes Understand normaization and database

More information

Dynamic Pricing Trade Market for Shared Resources in IIU Federated Cloud

Dynamic Pricing Trade Market for Shared Resources in IIU Federated Cloud Dynamic Pricing Trade Market or Shared Resources in IIU Federated Coud Tongrang Fan 1, Jian Liu 1, Feng Gao 1 1Schoo o Inormation Science and Technoogy, Shiiazhuang Tiedao University, Shiiazhuang, 543,

More information

Load Balance vs Energy Efficiency in Traffic Engineering: A Game Theoretical Perspective

Load Balance vs Energy Efficiency in Traffic Engineering: A Game Theoretical Perspective Load Baance vs Energy Efficiency in Traffic Engineering: A Game Theoretica Perspective Yangming Zhao, Sheng Wang, Shizhong Xu and Xiong Wang Schoo of Communication and Information Engineering University

More information

Finance 360 Problem Set #6 Solutions

Finance 360 Problem Set #6 Solutions Finance 360 Probem Set #6 Soutions 1) Suppose that you are the manager of an opera house. You have a constant margina cost of production equa to $50 (i.e. each additiona person in the theatre raises your

More information

NCH Software Warp Speed PC Tune-up Software

NCH Software Warp Speed PC Tune-up Software NCH Software Warp Speed PC Tune-up Software This user guide has been created for use with Warp Speed PC Tune-up Software Version 1.xx NCH Software Technica Support If you have difficuties using Warp Speed

More information

An FDD Wideband CDMA MAC Protocol for Wireless Multimedia Networks

An FDD Wideband CDMA MAC Protocol for Wireless Multimedia Networks An FDD ideband CDMA MAC Protoco for ireess Mutimedia Networks Xudong ang Broadband and ireess Networking (BN) Lab Schoo of Eectrica and Computer Engineering Georgia Institute of Technoogy Atanta, GA 3332

More information

500 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 31, NO. 3, MARCH 2013

500 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 31, NO. 3, MARCH 2013 500 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 3, NO. 3, MARCH 203 Cognitive Radio Network Duaity Agorithms for Utiity Maximization Liang Zheng Chee Wei Tan, Senior Member, IEEE Abstract We

More information

TCP/IP Gateways and Firewalls

TCP/IP Gateways and Firewalls Gateways and Firewas 1 Gateways and Firewas Prof. Jean-Yves Le Boudec Prof. Andrzej Duda ICA, EPFL CH-1015 Ecubens http://cawww.epf.ch Gateways and Firewas Firewas 2 o architecture separates hosts and

More information

A Branch-and-Price Algorithm for Parallel Machine Scheduling with Time Windows and Job Priorities

A Branch-and-Price Algorithm for Parallel Machine Scheduling with Time Windows and Job Priorities A Branch-and-Price Agorithm for Parae Machine Scheduing with Time Windows and Job Priorities Jonathan F. Bard, 1 Siwate Rojanasoonthon 2 1 Graduate Program in Operations Research and Industria Engineering,

More information

SELECTING THE SUITABLE ERP SYSTEM: A FUZZY AHP APPROACH. Ufuk Cebeci

SELECTING THE SUITABLE ERP SYSTEM: A FUZZY AHP APPROACH. Ufuk Cebeci SELECTING THE SUITABLE ERP SYSTEM: A FUZZY AHP APPROACH Ufuk Cebeci Department of Industria Engineering, Istanbu Technica University, Macka, Istanbu, Turkey - ufuk_cebeci@yahoo.com Abstract An Enterprise

More information

3.3 SOFTWARE RISK MANAGEMENT (SRM)

3.3 SOFTWARE RISK MANAGEMENT (SRM) 93 3.3 SOFTWARE RISK MANAGEMENT (SRM) Fig. 3.2 SRM is a process buit in five steps. The steps are: Identify Anayse Pan Track Resove The process is continuous in nature and handed dynamicay throughout ifecyce

More information

CONTRIBUTION OF INTERNAL AUDITING IN THE VALUE OF A NURSING UNIT WITHIN THREE YEARS

CONTRIBUTION OF INTERNAL AUDITING IN THE VALUE OF A NURSING UNIT WITHIN THREE YEARS Dehi Business Review X Vo. 4, No. 2, Juy - December 2003 CONTRIBUTION OF INTERNAL AUDITING IN THE VALUE OF A NURSING UNIT WITHIN THREE YEARS John N.. Var arvatsouakis atsouakis DURING the present time,

More information

Order-to-Cash Processes

Order-to-Cash Processes TMI170 ING info pat 2:Info pat.qxt 01/12/2008 09:25 Page 1 Section Two: Order-to-Cash Processes Gregory Cronie, Head Saes, Payments and Cash Management, ING O rder-to-cash and purchase-topay processes

More information

Take me to your leader! Online Optimization of Distributed Storage Configurations

Take me to your leader! Online Optimization of Distributed Storage Configurations Take me to your eader! Onine Optimization of Distributed Storage Configurations Artyom Sharov Aexander Shraer Arif Merchant Murray Stokey sharov@cs.technion.ac.i, {shraex, aamerchant, mstokey}@googe.com

More information

Chapter 3: e-business Integration Patterns

Chapter 3: e-business Integration Patterns Chapter 3: e-business Integration Patterns Page 1 of 9 Chapter 3: e-business Integration Patterns "Consistency is the ast refuge of the unimaginative." Oscar Wide In This Chapter What Are Integration Patterns?

More information

Leakage detection in water pipe networks using a Bayesian probabilistic framework

Leakage detection in water pipe networks using a Bayesian probabilistic framework Probabiistic Engineering Mechanics 18 (2003) 315 327 www.esevier.com/ocate/probengmech Leakage detection in water pipe networks using a Bayesian probabiistic framework Z. Pouakis, D. Vaougeorgis, C. Papadimitriou*

More information

subject to x dp = 1.0, d D (2a)

subject to x dp = 1.0, d D (2a) Lightpath Topology Configuration for Wavelength-routed IP/MPLS Networks for Time-Dependent Traffic Gaurav Agrawal, Deep Medhi Computer Science & Electrical Engineering Department University of Missouri

More information

Avaya Remote Feature Activation (RFA) User Guide

Avaya Remote Feature Activation (RFA) User Guide Avaya Remote Feature Activation (RFA) User Guide 03-300149 Issue 5.0 September 2007 2007 Avaya Inc. A Rights Reserved. Notice Whie reasonabe efforts were made to ensure that the information in this document

More information

CERTIFICATE COURSE ON CLIMATE CHANGE AND SUSTAINABILITY. Course Offered By: Indian Environmental Society

CERTIFICATE COURSE ON CLIMATE CHANGE AND SUSTAINABILITY. Course Offered By: Indian Environmental Society CERTIFICATE COURSE ON CLIMATE CHANGE AND SUSTAINABILITY Course Offered By: Indian Environmenta Society INTRODUCTION The Indian Environmenta Society (IES) a dynamic and fexibe organization with a goba vision

More information

SABRe B2.1: Design & Development. Supplier Briefing Pack.

SABRe B2.1: Design & Development. Supplier Briefing Pack. SABRe B2.1: Design & Deveopment. Suppier Briefing Pack. 2013 Ros-Royce pc The information in this document is the property of Ros-Royce pc and may not be copied or communicated to a third party, or used

More information

LADDER SAFETY Table of Contents

LADDER SAFETY Table of Contents Tabe of Contents SECTION 1. TRAINING PROGRAM INTRODUCTION..................3 Training Objectives...........................................3 Rationae for Training.........................................3

More information

FRAME BASED TEXTURE CLASSIFICATION BY CONSIDERING VARIOUS SPATIAL NEIGHBORHOODS. Karl Skretting and John Håkon Husøy

FRAME BASED TEXTURE CLASSIFICATION BY CONSIDERING VARIOUS SPATIAL NEIGHBORHOODS. Karl Skretting and John Håkon Husøy FRAME BASED TEXTURE CLASSIFICATION BY CONSIDERING VARIOUS SPATIAL NEIGHBORHOODS Kar Skretting and John Håkon Husøy University of Stavanger, Department of Eectrica and Computer Engineering N-4036 Stavanger,

More information

Early access to FAS payments for members in poor health

Early access to FAS payments for members in poor health Financia Assistance Scheme Eary access to FAS payments for members in poor heath Pension Protection Fund Protecting Peope s Futures The Financia Assistance Scheme is administered by the Pension Protection

More information

Program Management Seminar

Program Management Seminar Program Management Seminar Program Management Seminar The word s best management schoos are noted for their superior program execution, high eves of customer satisfaction, and continuous program improvements.

More information

A train dispatching model based on fuzzy passenger demand forecasting during holidays

A train dispatching model based on fuzzy passenger demand forecasting during holidays Journa of Industria Engineering and Management JIEM, 2013 6(1):320-335 Onine ISSN: 2013-0953 Print ISSN: 2013-8423 http://dx.doi.org/10.3926/jiem.699 A train dispatching mode based on fuzzy passenger demand

More information

arxiv:1506.05851v1 [cs.ai] 18 Jun 2015

arxiv:1506.05851v1 [cs.ai] 18 Jun 2015 Smart Pacing for Effective Onine Ad Campaign Optimization Jian Xu, Kuang-chih Lee, Wentong Li, Hang Qi, and Quan Lu Yahoo Inc. 7 First Avenue, Sunnyvae, Caifornia 9489 {xuian,kcee,wentong,hangqi,qu}@yahoo-inc.com

More information

3.5 Pendulum period. 2009-02-10 19:40:05 UTC / rev 4d4a39156f1e. g = 4π2 l T 2. g = 4π2 x1 m 4 s 2 = π 2 m s 2. 3.5 Pendulum period 68

3.5 Pendulum period. 2009-02-10 19:40:05 UTC / rev 4d4a39156f1e. g = 4π2 l T 2. g = 4π2 x1 m 4 s 2 = π 2 m s 2. 3.5 Pendulum period 68 68 68 3.5 Penduum period 68 3.5 Penduum period Is it coincidence that g, in units of meters per second squared, is 9.8, very cose to 2 9.87? Their proximity suggests a connection. Indeed, they are connected

More information

Advanced ColdFusion 4.0 Application Development - 3 - Server Clustering Using Bright Tiger

Advanced ColdFusion 4.0 Application Development - 3 - Server Clustering Using Bright Tiger Advanced CodFusion 4.0 Appication Deveopment - CH 3 - Server Custering Using Bri.. Page 1 of 7 [Figures are not incuded in this sampe chapter] Advanced CodFusion 4.0 Appication Deveopment - 3 - Server

More information

Human Capital & Human Resources Certificate Programs

Human Capital & Human Resources Certificate Programs MANAGEMENT CONCEPTS Human Capita & Human Resources Certificate Programs Programs to deveop functiona and strategic skis in: Human Capita // Human Resources ENROLL TODAY! Contract Hoder Contract GS-02F-0010J

More information

Ricoh Healthcare. Process Optimized. Healthcare Simplified.

Ricoh Healthcare. Process Optimized. Healthcare Simplified. Ricoh Heathcare Process Optimized. Heathcare Simpified. Rather than a destination that concudes with the eimination of a paper, the Paperess Maturity Roadmap is a continuous journey to strategicay remove

More information

Wide-Area Traffic Management for. Cloud Services

Wide-Area Traffic Management for. Cloud Services Wide-Area Traffic Management for Coud Services Joe Wenjie Jiang A Dissertation Presented to the Facuty of Princeton University in Candidacy for the Degree of Doctor of Phiosophy Recommended for Acceptance

More information

APPENDIX 10.1: SUBSTANTIVE AUDIT PROGRAMME FOR PRODUCTION WAGES: TROSTON PLC

APPENDIX 10.1: SUBSTANTIVE AUDIT PROGRAMME FOR PRODUCTION WAGES: TROSTON PLC Appendix 10.1: substantive audit programme for production wages: Troston pc 389 APPENDIX 10.1: SUBSTANTIVE AUDIT PROGRAMME FOR PRODUCTION WAGES: TROSTON PLC The detaied audit programme production wages

More information

WHITE PAPER BEsT PRAcTIcEs: PusHIng ExcEl BEyond ITs limits WITH InfoRmATIon optimization

WHITE PAPER BEsT PRAcTIcEs: PusHIng ExcEl BEyond ITs limits WITH InfoRmATIon optimization Best Practices: Pushing Exce Beyond Its Limits with Information Optimization WHITE Best Practices: Pushing Exce Beyond Its Limits with Information Optimization Executive Overview Microsoft Exce is the

More information

Precise assessment of partial discharge in underground MV/HV power cables and terminations

Precise assessment of partial discharge in underground MV/HV power cables and terminations QCM-C-PD-Survey Service Partia discharge monitoring for underground power cabes Precise assessment of partia discharge in underground MV/HV power cabes and terminations Highy accurate periodic PD survey

More information

eg Enterprise vs. a Big 4 Monitoring Soution: Comparing Tota Cost of Ownership Restricted Rights Legend The information contained in this document is confidentia and subject to change without notice. No

More information

READING A CREDIT REPORT

READING A CREDIT REPORT Name Date CHAPTER 6 STUDENT ACTIVITY SHEET READING A CREDIT REPORT Review the sampe credit report. Then search for a sampe credit report onine, print it off, and answer the questions beow. This activity

More information

NCH Software MoneyLine

NCH Software MoneyLine NCH Software MoneyLine This user guide has been created for use with MoneyLine Version 2.xx NCH Software Technica Support If you have difficuties using MoneyLine pease read the appicabe topic before requesting

More information

A New Statistical Approach to Network Anomaly Detection

A New Statistical Approach to Network Anomaly Detection A New Statistica Approach to Network Anomay Detection Christian Caegari, Sandrine Vaton 2, and Michee Pagano Dept of Information Engineering, University of Pisa, ITALY E-mai: {christiancaegari,mpagano}@ietunipiit

More information

Fixed income managers: evolution or revolution

Fixed income managers: evolution or revolution Fixed income managers: evoution or revoution Traditiona approaches to managing fixed interest funds rey on benchmarks that may not represent optima risk and return outcomes. New techniques based on separate

More information

Capacity of Multi-service Cellular Networks with Transmission-Rate Control: A Queueing Analysis

Capacity of Multi-service Cellular Networks with Transmission-Rate Control: A Queueing Analysis Capacity of Muti-service Ceuar Networs with Transmission-Rate Contro: A Queueing Anaysis Eitan Atman INRIA, BP93, 2004 Route des Lucioes, 06902 Sophia-Antipois, France aso CESIMO, Facutad de Ingeniería,

More information

Breakeven analysis and short-term decision making

Breakeven analysis and short-term decision making Chapter 20 Breakeven anaysis and short-term decision making REAL WORLD CASE This case study shows a typica situation in which management accounting can be hepfu. Read the case study now but ony attempt

More information

Budgeting Loans from the Social Fund

Budgeting Loans from the Social Fund Budgeting Loans from the Socia Fund tes sheet Pease read these notes carefuy. They expain the circumstances when a budgeting oan can be paid. Budgeting Loans You may be abe to get a Budgeting Loan if:

More information

Risk Margin for a Non-Life Insurance Run-Off

Risk Margin for a Non-Life Insurance Run-Off Risk Margin for a Non-Life Insurance Run-Off Mario V. Wüthrich, Pau Embrechts, Andreas Tsanakas February 2, 2011 Abstract For sovency purposes insurance companies need to cacuate so-caed best-estimate

More information

Bite-Size Steps to ITIL Success

Bite-Size Steps to ITIL Success 7 Bite-Size Steps to ITIL Success Pus making a Business Case for ITIL! Do you want to impement ITIL but don t know where to start? 7 Bite-Size Steps to ITIL Success can hep you to decide whether ITIL can

More information

Pay-on-delivery investing

Pay-on-delivery investing Pay-on-deivery investing EVOLVE INVESTment range 1 EVOLVE INVESTMENT RANGE EVOLVE INVESTMENT RANGE 2 Picture a word where you ony pay a company once they have deivered Imagine striking oi first, before

More information

A Similarity Search Scheme over Encrypted Cloud Images based on Secure Transformation

A Similarity Search Scheme over Encrypted Cloud Images based on Secure Transformation A Simiarity Search Scheme over Encrypted Coud Images based on Secure Transormation Zhihua Xia, Yi Zhu, Xingming Sun, and Jin Wang Jiangsu Engineering Center o Network Monitoring, Nanjing University o Inormation

More information

SNMP Reference Guide for Avaya Communication Manager

SNMP Reference Guide for Avaya Communication Manager SNMP Reference Guide for Avaya Communication Manager 03-602013 Issue 1.0 Feburary 2007 2006 Avaya Inc. A Rights Reserved. Notice Whie reasonabe efforts were made to ensure that the information in this

More information

Application-Aware Data Collection in Wireless Sensor Networks

Application-Aware Data Collection in Wireless Sensor Networks Appication-Aware Data Coection in Wireess Sensor Networks Xiaoin Fang *, Hong Gao *, Jianzhong Li *, and Yingshu Li +* * Schoo of Computer Science and Technoogy, Harbin Institute of Technoogy, Harbin,

More information

Leadership & Management Certificate Programs

Leadership & Management Certificate Programs MANAGEMENT CONCEPTS Leadership & Management Certificate Programs Programs to deveop expertise in: Anaytics // Leadership // Professiona Skis // Supervision ENROLL TODAY! Contract oder Contract GS-02F-0010J

More information

Discounted Cash Flow Analysis (aka Engineering Economy)

Discounted Cash Flow Analysis (aka Engineering Economy) Discounted Cash Fow Anaysis (aka Engineering Economy) Objective: To provide economic comparison of benefits and costs that occur over time Assumptions: Future benefits and costs can be predicted A Benefits,

More information

ILP Formulation and K-Shortest Path Heuristic for the RWA Problem with Allocation of Wavelength Converters

ILP Formulation and K-Shortest Path Heuristic for the RWA Problem with Allocation of Wavelength Converters ILP Formuation and K-Shortest Path Heuristic for the RWA Probem with Aocation of Waveength Converters Karcius D.R. Assis 1, A. C. B. Soares, Wiiam F. Giozza and Heio Wadman 1 UEFS- State University of

More information

Internal Control. Guidance for Directors on the Combined Code

Internal Control. Guidance for Directors on the Combined Code Interna Contro Guidance for Directors on the Combined Code ISBN 1 84152 010 1 Pubished by The Institute of Chartered Accountants in Engand & Waes Chartered Accountants Ha PO Box 433 Moorgate Pace London

More information

Design of Follow-Up Experiments for Improving Model Discrimination and Parameter Estimation

Design of Follow-Up Experiments for Improving Model Discrimination and Parameter Estimation Design of Foow-Up Experiments for Improving Mode Discrimination and Parameter Estimation Szu Hui Ng 1 Stephen E. Chick 2 Nationa University of Singapore, 10 Kent Ridge Crescent, Singapore 119260. Technoogy

More information

Network/Communicational Vulnerability

Network/Communicational Vulnerability Automated teer machines (ATMs) are a part of most of our ives. The major appea of these machines is convenience The ATM environment is changing and that change has serious ramifications for the security

More information

ELECTRONIC FUND TRANSFERS YOUR RIGHTS AND RESPONSIBILITIES

ELECTRONIC FUND TRANSFERS YOUR RIGHTS AND RESPONSIBILITIES About ELECTRONIC FUND TRANSFERS YOUR RIGHTS AND RESPONSIBILITIES The Eectronic Fund Transfers we are capabe of handing for consumers are indicated beow, some of which may not appy your account. Some of

More information

Mean-field Dynamics of Load-Balancing Networks with General Service Distributions

Mean-field Dynamics of Load-Balancing Networks with General Service Distributions Mean-fied Dynamics of Load-Baancing Networks with Genera Service Distributions Reza Aghajani 1, Xingjie Li 2, and Kavita Ramanan 1 1 Division of Appied Mathematics, Brown University, Providence, RI, USA.

More information

Market Design & Analysis for a P2P Backup System

Market Design & Analysis for a P2P Backup System Market Design & Anaysis for a P2P Backup System Sven Seuken Schoo of Engineering & Appied Sciences Harvard University, Cambridge, MA seuken@eecs.harvard.edu Denis Chares, Max Chickering, Sidd Puri Microsoft

More information

Art of Java Web Development By Neal Ford 624 pages US$44.95 Manning Publications, 2004 ISBN: 1-932394-06-0

Art of Java Web Development By Neal Ford 624 pages US$44.95 Manning Publications, 2004 ISBN: 1-932394-06-0 IEEE DISTRIBUTED SYSTEMS ONLINE 1541-4922 2005 Pubished by the IEEE Computer Society Vo. 6, No. 5; May 2005 Editor: Marcin Paprzycki, http://www.cs.okstate.edu/%7emarcin/ Book Reviews: Java Toos and Frameworks

More information

Life Contingencies Study Note for CAS Exam S. Tom Struppeck

Life Contingencies Study Note for CAS Exam S. Tom Struppeck Life Contingencies Study Note for CAS Eam S Tom Struppeck (Revised 9/19/2015) Introduction Life contingencies is a term used to describe surviva modes for human ives and resuting cash fows that start or

More information

Subject: Corns of En gineers and Bureau of Reclamation: Information on Potential Budgetarv Reductions for Fiscal Year 1998

Subject: Corns of En gineers and Bureau of Reclamation: Information on Potential Budgetarv Reductions for Fiscal Year 1998 GAO United States Genera Accounting Office Washington, D.C. 20548 Resources, Community, and Economic Deveopment Division B-276660 Apri 25, 1997 The Honorabe Pete V. Domenici Chairman The Honorabe Harry

More information

Vendor Performance Measurement Using Fuzzy Logic Controller

Vendor Performance Measurement Using Fuzzy Logic Controller The Journa of Mathematics and Computer Science Avaiabe onine at http://www.tjmcs.com The Journa of Mathematics and Computer Science Vo.2 No.2 (2011) 311-318 Performance Measurement Using Fuzzy Logic Controer

More information

Informatica PowerCenter

Informatica PowerCenter Brochure Informatica PowerCenter Benefits Support better business decisions with the right information at the right time Acceerate projects in days vs. months with improved staff productivity and coaboration

More information

Betting Strategies, Market Selection, and the Wisdom of Crowds

Betting Strategies, Market Selection, and the Wisdom of Crowds Betting Strategies, Market Seection, and the Wisdom of Crowds Wiemien Kets Northwestern University w-kets@keogg.northwestern.edu David M. Pennock Microsoft Research New York City dpennock@microsoft.com

More information

Modeling a Scenario-based Distribution Network Design Problem in a Physical Internet-enabled open Logistics Web

Modeling a Scenario-based Distribution Network Design Problem in a Physical Internet-enabled open Logistics Web 4 th Internationa conference on Information Systems, Logistics and Suppy Chain Quebec City August 26-29, 2012 Modeing a Scenario-based Distribution Network Design Probem in a Physica Internet-enabed open

More information

1B11 Operating Systems. Input/Output and Devices

1B11 Operating Systems. Input/Output and Devices University Coege London 1B11 Operating Systems Input/Output and s Prof. Steve R Wibur s.wibur@cs.uc.ac.uk Lecture Objectives How do the bits of the I/O story fit together? What is a device driver? 1B11-5

More information

Design and Analysis of a Hidden Peer-to-peer Backup Market

Design and Analysis of a Hidden Peer-to-peer Backup Market Design and Anaysis of a Hidden Peer-to-peer Backup Market Sven Seuken, Denis Chares, Max Chickering, Mary Czerwinski Kama Jain, David C. Parkes, Sidd Puri, and Desney Tan December, 2015 Abstract We present

More information

ELECTRONIC FUND TRANSFERS YOUR RIGHTS AND RESPONSIBILITIES. l l

ELECTRONIC FUND TRANSFERS YOUR RIGHTS AND RESPONSIBILITIES. l l ELECTRONIC FUND TRANSFERS YOUR RIGHTS AND RESPONSIBILITIES The Eectronic Fund Transfers we are capabe of handing for consumers are indicated beow some of which may not appy your account Some of these may

More information

Randomized Algorithms for Scheduling VMs in the Cloud

Randomized Algorithms for Scheduling VMs in the Cloud Randomized Agorithms for Scheduing VMs in the Coud Javad Ghaderi Coumbia University Abstract We consider the probem of scheduing VMs (Virtua Machines) in a muti-server system motivated by coud computing

More information

Federal Financial Management Certificate Program

Federal Financial Management Certificate Program MANAGEMENT CONCEPTS Federa Financia Management Certificate Program Training to hep you achieve the highest eve performance in: Accounting // Auditing // Budgeting // Financia Management ENROLL TODAY! Contract

More information

Iterative Water-filling for Load-balancing in Wireless LAN or Microcellular Networks

Iterative Water-filling for Load-balancing in Wireless LAN or Microcellular Networks terative Water-fiing for Load-baancing in Wireess LAN or Microceuar Networks Jeremy K. Chen Theodore S. Rappaport Gustavo de Veciana Wireess Networking and Communications Group (WNCG), be University of

More information

(12) Patent Application Publication (10) Pub. N0.: US 2006/0105797 A1 Marsan et al. (43) Pub. Date: May 18, 2006

(12) Patent Application Publication (10) Pub. N0.: US 2006/0105797 A1 Marsan et al. (43) Pub. Date: May 18, 2006 (19) United States US 20060105797A (12) Patent Appication Pubication (10) Pub. N0.: US 2006/0105797 A1 Marsan et a. (43) Pub. Date: (54) METHOD AND APPARATUS FOR (52) US. C...... 455/522 ADJUSTING A MOBILE

More information