Energy Efficient VM Scheduling for Cloud Data Centers: Exact allocation and migration algorithms

Size: px
Start display at page:

Download "Energy Efficient VM Scheduling for Cloud Data Centers: Exact allocation and migration algorithms"

Transcription

1 Energy Efficient VM Scheduling for Cloud Data Centers: Exact allocation and igration algoriths Chaia Ghribi, Makhlouf Hadji and Djaal Zeghlache Institut Mines-Téléco, Téléco SudParis UMR CNRS , Rue Charles Fourier 91011, Evry, France Eail: {chaia.ghribi, akhlouf.hadji, Abstract This paper presents two exact algoriths for energy efficient scheduling of virtual achines (VMs) in cloud data centers. Modeling of energy aware allocation and consolidation to iniize overall energy consuption leads us to the cobination of an optial allocation algorith with a consolidation algorith relying on igration of VMs at service departures. The optial allocation algorith is solved as a bin packing proble with a iniu power consuption objective. It is copared with an energy aware best fit algorith. The exact igration algorith results fro a linear and integer forulation of VM igration to adapt placeent when resources are released. The proposed igration is general and goes beyond the current state of the art by iniizing both the nuber of igrations needed for consolidation and energy consuption in a single algorith with a set of valid inequalities and conditions. Experiental results show the benefits of cobining the allocation and igration algoriths and deonstrate their ability to achieve significant energy savings while aintaining feasible convergence ties when copared with the best fit heuristic. Index Ters Energy efficiency, VM placeent, VM igration, Cloud data center, Linear integer prograing. I. INTRODUCTION Energy efficiency is becoing increasingly iportant for data centers and clouds. The wider adoption of cloud coputing and virtualization technologies has led to cluster sizes ranging fro hundreds to thousands of nodes for ini and large data centers respectively. This evolution induces a treendous rise of electricity consuption, escalating data center ownership costs and increasing carbon footprints. For these reasons, data centers now ebed onitoring capabilities and probes such as sart power distribution units (PDUs) to achieve energy efficiency and reduce overall cost. The fact that electricity consuption is set to rise 76% fro 2007 to 2030 [1] with data centers contributing an iportant portion of this increase ephasizes the iportance of reducing energy consuption in clouds. According to the Gartner Report [2], the average data center is estiated to consue as uch energy as households, and according to McKinsey report [3], The total estiated energy bill for data centers in 2010 is 11.5 billion and energy costs in a typical data center double every five years. Reducing the overall energy bill requires first and foreost energy and ecology conscious designs of data centers hardware and software, the use of solar and photovoltaic energy, relying on renewable energy sources and efficient cooling systes. The second and apparently ore odest contribution to energy efficient clouds is to introduce energy aware scheduling and placeent algoriths and enhanced resource anageent. In fact, cloud data centers are electricity guzzlers especially if resources are peranently switched on even if they are not used. An idle server consues about 70% of its peak power [4]. This waste of idle power is considered as a ajor cause of energy inefficiency. This paper is a contribution to the reduction of such excessive energy consuption using energy aware allocation and igration algoriths to have a axiu nuber of idle servers to put into sleep ode. Intel s Cloud Coputing 2015 Vision [5] stresses also the need for such dynaic resource scheduling approaches to iprove power efficiency of data centers by shutting down and putting to sleep idles servers. Our work proposes an exact energy aware allocation algorith using the classical forulation of the Bin-Packing proble. The ai of this algorith is to reduce the nuber of used servers or equivalently axiize the nuber of idle servers to put in sleep ode. To take into account workloads and service ties a linear integer prograing algorith is used to optiize constantly the nuber of used servers after service departures. This igration algorith is cobined with the exact allocation algorith to reduce overall energy consuption in the data centers. The proposed algoriths act as an energy consuption aware VM scheduler and can be used to enhance current infrastructure anagers and schedulers such as OpenNebula [6] and OpenStack [7]. The power consuption indicators can be provided by energy consuption estiation tools such as jouleeter [8]. A dedicated siulator is used to assess perforance and crosscheck with the perforance results produced by the exact algoriths. Evaluation results show that the exact allocation algorith cobined with igration reduces considerably the nuber of required servers to serve a given load and can thus iniize power consuption in data centers. Section II of this paper describes a related work addressing consolidation using different igration techniques. The syste odel and the derived algoriths are addressed in Section III. Perforance evaluation is the object of section IV. II. RELATED WORK Authors in [9] consolidate applications or tasks on reduced nuber of physical achines to switch off achines in

2 2 surplus. This approach differs fro ours because consolidation is achieved at the task level, rather than the virtual achine (VM) level (or IaaS level)in our case, and hence fits better the Platfor or Software as a Service (PaaS, SaaS) levels. Allocation or placeent is also static as opposed to our dynaic placeent according to workload where we apply igration to allocate and reallocate VMs. Reference [10] addresses policies for dynaic VMs reallocation using VMs igration according to CPU perforance requireents. Their ost effective policy, a double threshold policy, is based on the idea of setting upper and lower utilization thresholds for hosts and keeping the total utilization of the CPU of all the VMs between these thresholds. If the CPU utilization of a host exceeds the upper threshold, soe VMs are igrated and if it falls below the lower threshold, all the hosted VMs should be igrated. One of the closest work to our paper is found in [11]. Authors treat the proble of consolidating VMs in a server by igrating VMs with steady and stable capacity needs. They proposed an exact forulation based on a linear progra described by a too sall nuber of valid inequalities. Indeed, this description does not allow solving, in reasonable tie and in an optial way, probles involving allocation of a large nuber of ites (or VMs) to any bins (or Servers). In order to scale and find solutions for large sizes, the authors resorted to a heuristic using a static and a dynaic consolidation of VMs to reduce energy consuption of the hosting nodes or servers. In [12], authors presented a server consolidation (Sercon) algorith which consists of iniizing the nuber of used nodes in a data center and iniizing the nuber of igrations at the sae tie. They copared their algorith with the well known placeent heuristic FFD (First-Fit Decreasing) [13], to solve the Bin-Packing proble, and have shown the efficiency of Sercon to consolidate VMs and iniize igrations. However, Sercon is a heuristic that can not always reach or find the optial solution. Our proposed algorith is based on an exact forulation of the consolidation proble with igrations. We anage to optially consolidate VMs in servers while iniizing the energy cost of igrations. In [14], authors presented an approach EnaCloud for dynaic live placeent taking into account energy efficiency in a cloud platfor. They proposed an energy-aware heuristic algorith in order to save energy by iniizing the nuber of running servers. Another study relying on dynaic resource allocation is presented in [15]. The authors presented a natureinspired VM consolidation algorith inspired fro an Ant Colony Optiization. This algorith ais at reducing the nuber of used physical achines and thus saves energy. III. THE SYSTEM MODEL The odel considers infrastructure providers allocating physical resources instances to host users and tenants applications or equivalently for this paper VMs. The physical resources are seen as servers. It is assued that applications are packaged into virtual achines to be hosted by the infrastructure providers. The cloud providers save energy and reduce power consuption by packing and consolidating through igration of VMS to axiize the nuber of idle servers to put to sleep ode. Figure 1 depicts the syste odel coposed of the proposed energy efficient allocation and igration algoriths contributing to scheduling, an energy consuption estiator and cloud anagers that handle infrastructure resource instantiation and anageent. Each odule is briefly described to set the stage for the analytical odeling of the energy efficient resource allocation proble in clouds. Cloud IaaS anager (e.g. OpenStack [7], OpenNebula [6] and Eucalyptus [16]) control and anage cloud resources and handle clients requests, VM scheduling and fetch and store iages in storage spaces. Energy estiation odule is an interediate odule between the cloud infrastructure anager and the energyaware scheduler. The odule can rely for exaple on an energy estiation tool as in Jouleeter [8] that uses power odels to infer power consuption of VMs or servers fro resource usage. Energy-aware VM scheduler responsible for the energy aware VM placeent in the data center is the focus of our energy consuption optiization odel. This green scheduler is basically coposed of two odules. An allocation odule and a igration odule. The role of the allocation odule is to perfor the initial VM placeent using our exact VM allocation algorith. The dynaic consolidation of virtual achines is handled by the igration odule that iniizes the nuber of used or activated servers thanks to our exact VM igration algorith. The unused servers are shut down or put into sleep ode. All the needed inforation (servers and VMs) to run the algoriths are retrieved via the Cloud IaaS anager that is also used to execute the VM deployent and igration actions. To derive the syste odel, we consider the size n of client requests in ters of the nuber of required VMs and the types of desired VM instances (e.g., sall, ediu, large). Each V M i is characterized by a lifetie t i and a axiu power consuption p i. Each server or hosting node j, fro the data center, has a power consuption liit or power cap noted. This can be fixed by cloud adinistrators. We assue that all servers are hoogeneous; extending the odel to heterogeneous servers is trivial but will increase coplexity and will not necessarily provide additional insight. The approach adopted to achieve energy efficiency in our proposal is to use a bin packing algorith for optial placeent of user requests and to follow with dynaic consolidation once a sufficient nuber of departures have occurred. The dynaic consolidation is handled by the igration algorith which regroups VMs to free as any servers as possible to put the into sleep ode or to shut the down. A. Exact VM Allocation Algorith The proposed exact VM allocation algorith is an extended Bin-Packing approach through the inclusion of valid conditions expressed in the for of constraints or inequalities. The objective is to pack ites (VMs in our case) into a set of bins (servers or nodes hosting the VMs) characterized by their power consuptions. In addition to n, the nuber

3 3 Fig. 1. The syste odel of requested VMs, We define the nuber of servers,, available in the data center. The servers are assued to have the sae power consuption liit:, {j = 1, 2,..., }. At run-tie, each server j hosting a nuber of VMs is characterized by its current power consuption: P j,current. Since the objective is to iniize the energy consuption of the data centers, we define as key decision variable e j for each server j that is set to 1 if server j is selected to host VMs, 0 if it is not selected. In addition, we define the bivalent variable x ij to indicate that V M i has been placed in server j and set x ij to 1; x ij = 0 otherwise. The objective function to place all the deands (or VMs) in a iniu nuber of servers can be expressed using: in Z = e j (1) This optiization is subject to a nuber of linear constraints reflecting the capacity liits of the servers and obvious facts such as a VM can only be assigned to one server or a server can only host VMs according to the aount of reaining resources: 1) Each server has a power liit that cannot be exceeded when serving or hosting VMs and this occurs according to reaining capacity: n p i x ij e j P j,current, j = 1,..., (2) 2) A cloud provider has to fulfil all requests within a prescribed SLA or quota and each requested VM will be assigned to one and only one server: x ij = 1, i = 1,..., n (3) 3) For servers verifying the condition > P j,current and P j,current 0, the total nuber of used servers is lower bounded by Pj,current. This adds he following inequality to the odel: e j P j,current (4) The exact and extended Bin-Packing VM allocation odel can be suarized by luping the objective function with all the constraints and conditions and constraints into the following set of equations: in Z = e j (5) Subject To: n p i x ij e j P j,current, j = 1,..., (6) x ij = 1, i = 1,..., n (7) e j P j,current { 1, if the server j is used; e j = 0, otherwise. { 1, if the V Mi is placed in server j; x ij = 0, otherwise. (8) (9) (10) All the variables and constants used in the odel are listed for easy reference below: n is the size of the request in nuber of requested VMs. is the nuber of servers in the data center. p i represents the power consuption of V M i. x ij is a bivalent variable indicating that V M i is assigned to a server j. e j is a variable used to indicate whether the server j is used or not. represents the axiu power consuption of server j. P j,current represents the current power consuption of server j (P j,current = P j,idle + k p k with VM k hosted by server j). P j,idle represents the power consuption of server j when it is idle. Constraints in server CPU, eory and storage are also added to the odel to confine even further the odel convex hull: n cpu i x ij CP U j e j (11) where cpu i is the requested CPU by V M i. CP U j is the CPU capacity of server j. e i x ij MEM j e j (12) where e i is the requested eory by V M i and MEM j is the eory capacity of server j. sto i x ij ST O j e j (13) where sto i is the requested storage by V M i and ST O j is the storage capacity of server j.

4 4 In this paper we assue that these constraints are et and verified and we hence only need to focus on the energy efficiency constraints through (2). B. Best-Fit forulation and heuristic algorith The exact and extended Bin-Packing is copared to a Best- Fit heuristic adaptation of the well known Best-Fit algorith [13]. The heuristic proposed to achieve energy efficient VM placeent consists of two steps: 1) sorting the requested VMs in decreasing order of power consuption. This builds soehow an ordered stack that is used in the second step for packing VMs in available servers; 2) The sorted VMs are handled starting fro the top of the stack and attepting to place the ost power consuing VMs in the server with the sallest reaining power consuption budget until a VM down the stack fits in this target server. The process repeats or continues until all VMs in the stack are placed and packed as uch as possible in the ost occupied servers. This will tend to free servers for sleep ode or switching off. As this Best-Fit heuristic algorith tries to approxiate the Bin-Packing algorith, it is selected for coparison with our exact VM allocation proposal. The allocation algoriths are cobined with a igration algorith to iniize overall data center power consuption. In our case, the objective is to benchark the exact VM allocation and igration algoriths with at least one well known heuristic algorith. The Best- Fit heuristic was selected since it is known to achieve good suboptial perforance copared with classical Bin-Packing. C. Exact VMs Migration The placed and running VMs in the servers will gradually leave the syste as their related jobs end. These departures are the opportunity to re-optiize the placeent by igrating VMs always in the syste for consolidation in a iniu nuber of fully packed severs. A igration algorith based on an integer linear progra (ILP) is presented to achieve the consolidation. This ILP algorith consists in introducing a nuber of valid inequalities to reduce the span of the convex hull of the igration proble. The atheatical odel for the VM consolidation via igration relies on a linear integer prograing forulation. The objective for the algorith is to igrate VMs fro nodes selected as source nodes (those the algorith ais at eptying so they can be turned off) to other selected destination nodes (those the algorith ais at filling so they serve a axiu nuber of VMs within their capacity liits). Ideally, the algorith should iniize the nuber of active nodes, axiize the overall nuber of VMs handled by the active nodes and hence axiize the nuber of unused, epty or idle nodes. The algorith should also iniize the power consuption caused by igrations. If the power consuption or cost of VM igration is unifor or hoogeneous across hosting nodes or servers, the objective reduces to iniizing the nuber of igrations. The igration concerns the set of non idle servers, <, whose power consuptions are lower than with j in. Despite the slight reduction in size <, the proble reains NP-hard. Hence, we resort to an exact algorith based on linear integer prograing to address optial igration for practical proble sizes or nuber of instances. Fig. 2. Exaple of VMs igration The objective function for the optial VM igration and consolidation can be expressed as the axiization of the nuber of idle servers in the infrastructure: q i ax M = P i,idle y i p kz ijk (14) k=1 where y i = 1 is used to indicate that server i is idle and y i = 0 eans that at least one VM is active in server i. P i,idle is the power consued by idle servers, p k is the cost in ters of consued power when igrating V M k. A fixed power consuption p k is assigned to each VM instance according to instance type to differentiate between sall, ediu and large VMs. Variable z ijk is the bivalent variable expressing igration of V M k fro server i to server j. Variable q i is the total nuber of VMs hosted on server i and that are candidate for igration into destination servers, especially server j in equation (14). The objective function (14) is subject to the igration constraints cited earlier. These conditions are forally expressed through valid inequalities and constraints that have to be respected when iniizing overall energy consuption. 1) When igrating V M k fro a server i to a server j (see figure 3), the algorith ust prevent backward igrations and can only igrate into one specific destination node. Stated in an equivalent way: if a V M k is igrated fro a server i (source) to a server j (destination), it can not be igrated to any other server l (l j). The proposed inequality (15) also ensures that VMs in destination node and VMs igrated to destination nodes are not igrated as we are aiing at filling these nodes instead of eptying the obviously. This is reflected by the inequality : z ijk + z jlk 1; (15) 2) To strengthen further the previous condition, a valid inequality is added to ensure that when a V M k is igrated fro server i to server j (see figure 4), igration to other nodes l (l j) are prevented or forbidden:,j i z ijk 1; (16)

5 5 Fig. 3. A server candidate to a igration should not igrate his own VMs z ijk + z jlk 1 (22) i = 1,...,, j = 1,...,, k = 1,..., q i, k = 1,..., q j, j i, and l = 1,...,, l j, k k.,j i z ijk 1 (23) i = 1,...,, j = 1,...,, k = 1,..., q i, l = 1,...,, l j. q i p k z ijk ( P j,current ) (1 y j ) (24) k=1 j = 1,...,, j i Fig. 4. A V M k can not be igrated to any servers in the sae tie 3) A server j is liited by its power consuption liit. The inequality (17) allows each server j to host VMs without exceeding its power liit: q i p k z ijk ( P j,current ) (1 y j ) (17) k=1 Where P j,current is the current power consuption of server j. 4) If a non-idle server i is a source of VM igration, then it should igrate all of its hosted VMs in order to be put to sleep ode or shut down once copletely eptied: q i z ijk = q i y i, i = 1,...,, j i (18) k=1 5) Another valid inequality is the upper bound in the total nuber of epty servers: y i P j,current (19) 6) Another iportant aspect is to avoid igration of VMs whose lifetie or leftover lifetie t k is shorter than the tie needed to ake igration decisions T 0 : z ijk t k T 0, (20) where t k = t k CurrentT ie, where CurrentT ie represents current or VM igration handling tie. The optial VM consolidation and igration odel and objective function (14) can be suarized for convenience with all the valid conditions as: Subject To: q i ax M = P i,idle y i p kz ijk (21) k=1 q i z ijk = q i y i, i = 1,...,, j i (25) k=1 y i P j,current (26) z ijk t k T 0, (27) { 1, if the V Mk is igrated fro a server i to a server j; z ijk = 0, otherwise. (28) { 1, if the Server i is idle; y i = (29) 0, otherwise. D. Cobination of allocation and igration algoriths Figure 5 suarizes how the allocation and igration algoriths are cobined to achieve inial energy consuption in infrastructure nodes and hence data centers. Both the exact Bin-Packing extension and the Best-Fit heuristic are used to ensure optial and suboptial placeent respectively. Recall that the two algoriths allow us to cross check their relative perforance and benchark the odified Best-Fit with the proposed exact allocation solution since the algoriths exhibit different optiality and convergence tie characteristics. Upon arrival of VM placeent and resource requests these two algoriths select the ost appropriate nodes to host new VMs in available and active nodes. Whenever necessary these algoriths ay resort to turning new nodes on, when the set of active nodes are full and cannot host the new arriving VM instances (sall, ediu or large). Both algoriths will attept serving the requests in the currently active nodes and will of course typically avoid turning any new nodes on. The algoriths are cobined with the igration algorith that is launched if a nuber of VM jobs terinate since their dedicated resources becoe available for opportunistic reuse and for ore efficient resource allocation and distribution. These departures are the opportunity for the consolidation algorith to rearrange allocations by oving VMs into the

6 6 Fig. 5. Cobination of the igration algorith with the two allocation algoriths SP ECcpu2006 [20] workloads (471.onetpp, 470.lb and 445.gobk) were evaluated. Estiated power consuptions were found to be between 25 and 28 watts for these eleents. Without loss of generality and to ease intuitive verification of the results, we refer to these published consuption to associate to each VM type (sall, ediu and large) an energy consuption p i respectively equal to 10 watts (low), 20 watts (ediu) and 30 watts (high) to stay in line with published values. The requests for resources to serve VMs have a constant arrival rate. The requested VM instance types are discrete unifor in [1, 3] (1 for sall, 2 for ediu and 3 for large instances). The VM sizes are arbitrarily drawn as unifor in [1, 3] and classified according to their type. We retained only the rando drawings that fulfill the VM characteristics in size and type. sallest possible set of nodes. All eptied or freed servers (or nodes) are turned off to iniize energy consuption. The consolidation is achieved by the exact igration algorith that oves VMs fro selected source nodes to selected destination nodes. The end result is the activation and use of the sallest set of nodes in the data centers. IV. NUMERICAL RESULTS Our proposed algoriths are evaluated through a Java language ipleentation and the linear solver CPLEX [17]. A dedicated siulator is developed to conduct the perforance assessents and the coparison. The objective of the nuerical evaluation is to quantify the percentage of energy savings or power consuption savings that can be expected when cobining the exact allocation algorith and the consolidation process using our proposed exact igration algorith. The answers provided by the nuerical analysis concern also the scalability and coplexity of the proposed algoriths in the size of the data centers and the arrival rate of requests for resources to host VMs which is also synonyous to load on the syste. Note, however, that the siulation are conducted for an arrival rate strictly lower than the rate of VM job departures fro the syste; thus siulations correspond to cases where the likelihood of finding an optial or a good solution is high. The assessent scenarios correspond to data centers with 100 servers or nodes for the first five experients. In the last two experients 200 servers are considered. We collect essentially as perforance indicators, the percentage of used servers (which autoatically provides the energy consued or saved by the algoriths) and the tie required for the algoriths to find their best solutions (optial for the exact algoriths). All the servers have a power consuption cap set to 200 watts (the peak power of a typical server is around 250 watts [18]). To perfor per-vm power estiation we referred to a power estiation odel proposed in [19]. Three SP ECcpu2006 [20] workloads (454.calculix, 482.sphinx and 435.groacs) with high, ediu and low power consuption were considered. Their associated power consuption is close to 13 watts, 11 watts and 10 watts respectively. The power estiation odel proposed in [8] provided additional insight. The power consuption of other Fig. 6. Coparison between the exact and heuristic allocation algoriths Figure 6 depicts results of a coparison between the adapted Best-Fit heuristic and our exact extended Bin- Packing allocation algoriths. The siulations correspond to 100 servers and resource requests in nuber of VMs in the [1, 200] range. The lifetie of the VMs are unifor in [30s, 180s]. That is VM jobs last at least 30s and will terinate in less than 180s. The exact allocation algorith as expected outperfors the Best-Fit heuristic for the 1000s tie interval siulated and reported in Figure 6. The Best-Fit heuristic uses ore often all available nodes or servers (100% ordinate value in Figure 6) while the exact algorith anages to use fewer nodes with 10 to 50% ore unused servers that Best-Fit. Figure 7 extends the analysis for the exact and extended Bin- Packing allocation algorith by coparing its perforance with and without consolidation. When the exact algorith is cobined with the igration algorith (that uses igration to epty soe nodes) it can significantly provide additional power savings or energy consuption reduction. The average gain can be estiated to be 10 to 20% ore servers that could be turned off. The average line for the exact algorith is around 80% of servers used while the average for the exact algorith with igration is ore in the order of 60%. Figure 8 pursues the analysis for the exact bin-packing VM allocation algorith by reporting perforance as a function

7 7 Figures 9, 10 and 11 address the perforance of the consolidation algorith via the analysis of the tie needed to achieve igrations of VMs fro source nodes to destination nodes in order to free as any servers as possible and gain the opportunity to shut the down. The assessent is perfored consequently on the active servers, i.e. those currently serving VMs and candidate for consolidation. The perforance as a function of increasing nuber of active nodes to consolidate is reported in the three figures for = 5, = 10 and = 20. For = 5, consolidation after igration fro source to destination nodes can be achieved in the illiseconds tie scales (few to 300 s in Figure 9). The nuber of hosted VMs to consolidate varies fro 5 to 30 for this siulation. Fig. 9. Execution tie of the exact igration algorith ( = 5) Fig. 7. Perforance coparison of the exact allocation algorith with and without igration of data center sizes and VM requests induced load. The tie before convergence to the optial placeent is reported as a function of data center size (fro 100 to 1000 nodes or servers) for request sizes ranging fro 50 to 500 VMs. Clearly, because the proble is NP-Hard, the convergence tie of the exact algorith grows exponentially for requests exceeding 300 VMs; especially for nuber of servers beyond 400. The tie needed to find the optial solutions reains acceptable and reasonable, within 10 s, for data center sizes below 500 receiving requests less than 400 VMs. The tie needed for convergence grows unacceptably high outside of this operating range for the siulated scenarios (tens of seconds to few inutes). This otivated the use of the Best-Fit algorith to find solutions faster even if they are bound to be suboptial as reported in Figure 6. The tie needed for consolidation increases to seconds in Figure 10 for = 10. The curve is reported for up to 60 VMs hosted in the nodes considered or subject to consolidation/igration. When the nuber of servers to consolidate increases further, as shown in Figure 11 for = 20, the convergence ties ove to orders of tens to hundreds of inutes (for the extree case, on the curve upper right corner, this reaches 180 inutes for 120 hosted VMs). These three figures highlight the liits of the exact igration algorith with increasing nuber of servers to consolidate. Fig. 10. Execution tie of the exact igration algorith ( = 10) Fig. 8. Execution tie of the Exact Allocation Algorith Fig. 11. Execution tie of the exact igration algorith ( = 20) The next experients and siulations address the achievable energy savings using different VM requests inter-arrival ties (noted by λ 1 ) and lifeties (represented by µ 1 that also reflects the service rate µ 1 ) in order to assess the perforance for variable syste loads since the ration λ/µ governs perforance. The nuber of servers has been fixed to 200 hosting nodes for the reported results in Table I. One hundred (100) siulation runs are average for each paraeter setting in the table. Table I reports the energy savings with

8 8 the igration algorith copared to the allocation algorith without igration. TABLE I TABLE OF PERCENTAGE OF GAINED ENERGY WHEN MIGRATION IS USED λ µ 1 (s) (s) ,55 36,59 00,00 00,00 00,00 00, ,29 34,00 35,23 38,50 00,00 00, ,48 27,39 35,21 40,32 41,89 36, ,77 18,85 22,02 32,31 39,90 40, ,86 16,17 19,85 22,30 39,20 36, ,63 14,29 18,01 22,13 25,15 30, ,10 14,00 14,86 15,90 22,91 23, ,01 10,20 10,91 15,34 17,02 21, ,80 09,52 10,31 14,70 16,97 19, ,90 07,50 08,40 12,90 16,00 14,97 Energy savings depend evidently on the service rate or the lifetie of VMs or the duration of their jobs relative to the load induced by the VM resource requests. Savings in the siulation can reach as high as 41.89% for inter-arrival ties of 25 seconds and job durations of 30 seconds. For less favorable ratios or loads, the savings for the scenarios tested in the evaluation are less significant but reain respectable (5.90% for the highest loads (inter-arrivals tie is equal to 5seconds) and longer job durations (of 100sec). Fig. 12. Energy savings In order to coplete the analysis, the energy savings that can be achieved by the Best-Fit, the exact allocation and the exact allocation cobined with igration are copared for siilar scenarios with a restricted set of paraeter settings (λ 1 = 10s). All the servers are initially considered OFF, which eans that the energy saving is initialized to 100%. Figure 12 depicts the evolution of the percentage of energy saved by the three algoriths. The obvious dependence on syste load is reflected by the gradual decrease of energy savings for increasing VM lifeties (or increasing job durations). For the exact Bin-Packing allocation algorith the energy savings reain quite high at low load; 90% of energy savings for the exact allocation only and 95% cobined igration. Energy savings achieved by the Best-Fit heuristic stay below the percentages achieved by the exact allocation algorith with or without igration at all different loads. As usual the gap will vanish for saturated systes as the servers will be always busy and peranently turned on, the exact algorith can nevertheless provide gains if intelligent groupings and rearrangeents are applied (this is however out of scope of this paper). V. CONCLUSIONS This paper explored the proble of VM placeent in cloud providers data centers, which is known as NP-hard. Our original contribution consists in using exact algoriths for placeent and consolidation that iniize jointly energy consuption and igration costs in acceptable convergence ties copared slightly faster heuristics such as best fit. We have proposed a linear integer progra corresponding to an exact allocation algorith coupled with an efficient exact VM igration algorith to reduce energy consuption via consolidation. Significant energy savings can be obtained depending on syste loads. At low load the gains can be quite significant. These gains at high loads even if uch lower reain sufficiently valuable using the proposed algoriths. REFERENCES [1] World Energy Outlook 2009 FACT SHEET. [2] Gartner report, financial ties, [3] J. Kaplan, W. Forrest, and N. Kindler, Revolutionizing data center energy efficiency, McKinsey Copany Tech Rep, p. 15, July [4] E. Naone, Conjuring clouds, Technology Review, vol. 112, no. 4, pp , [5] Intel s cloud coputing 2015 vision. [6] Open nebula. [7] OpenStack. [8] A. Kansal, F. Zhao, J. Liu, N. Kothari, and A. A. Bhattacharya, Virtual achine power etering and provisioning, in Proceedings of the 1st ACM syposiu on Cloud coputing, ser. SoCC 10. New York, NY, USA: ACM, 2010, pp [Online]. Available: [9] S. Srikantaiah, A. Kansal, and F. Zhao, Energy aware consolidation for cloud coputing, in Proceedings of the 2008 conference on Power aware coputing and systes, ser. HotPower 08. Berkeley, CA, USA: USENIX Association, 2008, pp [Online]. Available: [10] A. Beloglazov and R. Buyya, Energy efficient resource anageent in virtualized cloud data centers, in Proceedings of the th IEEE/ACM International Conference on Cluster, Cloud and Grid Coputing, ser. CCGRID 10. Washington, DC, USA: IEEE Coputer Society, 2010, pp [Online]. Available: [11] T. C. Ferreto, M. A. S. Netto, R. N. Calheiros, and C. A. F. De Rose, Server consolidation with igration control for virtualized data centers, Future Generation Coputer Systes, vol. 27, no. 8, pp , [Online]. Available: retrieve/pii/s x [12] A. Murtazaev and S. Oh, Sercon : Server consolidation algorith using live igration of virtual achines for green coputing, Iete Technical Review, vol. 28, no. 3, pp , [Online]. Available: inward/record.url?eid=2-s &partnerid=40&d5= c44ad c05756f92cb667af7374 [13] E. G. Coffan, J. Csirik, and G. Woeginger, Approxiate Solutions to Bin Packing Probles. Oxford University Press, [14] B. Li, J. Li, J. Huai, T. Wo, Q. Li, and L. Zhong, Enacloud: An energy-saving application live placeent approach for cloud coputing environents, in Proceedings of the 2009 IEEE International Conference on Cloud Coputing, ser. CLOUD 09. Washington, DC, USA: IEEE Coputer Society, 2009, pp [Online]. Available: [15] L. Chiakurthi and M. K. SD, Power efficient resource allocation for clouds using ant colony fraework, CoRR, vol. abs/ , [16] Eucalyptus. [17] IBM ILOG CPLEX (August 2012). solver-for-linear-prograing2?gclid=clki- 56wrbACFVMetAodGRPLVA. [18] Green IT: Making the Business Case, Cognizant Report, January [19] R. B. Yol, X. Martorell, J. Torres, and E. Ayguade, Technical report upc-dac-rr-cap-18 accurate energy accounting for shared virtualized environents using pc-based power odeling techniques. [20] J. L. Henning, Spec cpu2006 benchark descriptions, SIGARCH Coputer Architecture News, vol. 34, no. 4, pp. 1 17, 2006.

The Research of Measuring Approach and Energy Efficiency for Hadoop Periodic Jobs

The Research of Measuring Approach and Energy Efficiency for Hadoop Periodic Jobs Send Orders for Reprints to reprints@benthascience.ae 206 The Open Fuels & Energy Science Journal, 2015, 8, 206-210 Open Access The Research of Measuring Approach and Energy Efficiency for Hadoop Periodic

More information

An Innovate Dynamic Load Balancing Algorithm Based on Task

An Innovate Dynamic Load Balancing Algorithm Based on Task An Innovate Dynaic Load Balancing Algorith Based on Task Classification Hong-bin Wang,,a, Zhi-yi Fang, b, Guan-nan Qu,*,c, Xiao-dan Ren,d College of Coputer Science and Technology, Jilin University, Changchun

More information

Exploiting Hardware Heterogeneity within the Same Instance Type of Amazon EC2

Exploiting Hardware Heterogeneity within the Same Instance Type of Amazon EC2 Exploiting Hardware Heterogeneity within the Sae Instance Type of Aazon EC2 Zhonghong Ou, Hao Zhuang, Jukka K. Nurinen, Antti Ylä-Jääski, Pan Hui Aalto University, Finland; Deutsch Teleko Laboratories,

More information

Dynamic Placement for Clustered Web Applications

Dynamic Placement for Clustered Web Applications Dynaic laceent for Clustered Web Applications A. Karve, T. Kibrel, G. acifici, M. Spreitzer, M. Steinder, M. Sviridenko, and A. Tantawi IBM T.J. Watson Research Center {karve,kibrel,giovanni,spreitz,steinder,sviri,tantawi}@us.ib.co

More information

Cooperative Caching for Adaptive Bit Rate Streaming in Content Delivery Networks

Cooperative Caching for Adaptive Bit Rate Streaming in Content Delivery Networks Cooperative Caching for Adaptive Bit Rate Streaing in Content Delivery Networs Phuong Luu Vo Departent of Coputer Science and Engineering, International University - VNUHCM, Vietna vtlphuong@hciu.edu.vn

More information

ASIC Design Project Management Supported by Multi Agent Simulation

ASIC Design Project Management Supported by Multi Agent Simulation ASIC Design Project Manageent Supported by Multi Agent Siulation Jana Blaschke, Christian Sebeke, Wolfgang Rosenstiel Abstract The coplexity of Application Specific Integrated Circuits (ASICs) is continuously

More information

A Scalable Application Placement Controller for Enterprise Data Centers

A Scalable Application Placement Controller for Enterprise Data Centers W WWW 7 / Track: Perforance and Scalability A Scalable Application Placeent Controller for Enterprise Data Centers Chunqiang Tang, Malgorzata Steinder, Michael Spreitzer, and Giovanni Pacifici IBM T.J.

More information

PERFORMANCE METRICS FOR THE IT SERVICES PORTFOLIO

PERFORMANCE METRICS FOR THE IT SERVICES PORTFOLIO Bulletin of the Transilvania University of Braşov Series I: Engineering Sciences Vol. 4 (53) No. - 0 PERFORMANCE METRICS FOR THE IT SERVICES PORTFOLIO V. CAZACU I. SZÉKELY F. SANDU 3 T. BĂLAN Abstract:

More information

Extended-Horizon Analysis of Pressure Sensitivities for Leak Detection in Water Distribution Networks: Application to the Barcelona Network

Extended-Horizon Analysis of Pressure Sensitivities for Leak Detection in Water Distribution Networks: Application to the Barcelona Network 2013 European Control Conference (ECC) July 17-19, 2013, Zürich, Switzerland. Extended-Horizon Analysis of Pressure Sensitivities for Leak Detection in Water Distribution Networks: Application to the Barcelona

More information

Real Time Target Tracking with Binary Sensor Networks and Parallel Computing

Real Time Target Tracking with Binary Sensor Networks and Parallel Computing Real Tie Target Tracking with Binary Sensor Networks and Parallel Coputing Hong Lin, John Rushing, Sara J. Graves, Steve Tanner, and Evans Criswell Abstract A parallel real tie data fusion and target tracking

More information

Online Bagging and Boosting

Online Bagging and Boosting Abstract Bagging and boosting are two of the ost well-known enseble learning ethods due to their theoretical perforance guarantees and strong experiental results. However, these algoriths have been used

More information

International Journal of Management & Information Systems First Quarter 2012 Volume 16, Number 1

International Journal of Management & Information Systems First Quarter 2012 Volume 16, Number 1 International Journal of Manageent & Inforation Systes First Quarter 2012 Volue 16, Nuber 1 Proposal And Effectiveness Of A Highly Copelling Direct Mail Method - Establishent And Deployent Of PMOS-DM Hisatoshi

More information

This paper studies a rental firm that offers reusable products to price- and quality-of-service sensitive

This paper studies a rental firm that offers reusable products to price- and quality-of-service sensitive MANUFACTURING & SERVICE OPERATIONS MANAGEMENT Vol., No. 3, Suer 28, pp. 429 447 issn 523-464 eissn 526-5498 8 3 429 infors doi.287/so.7.8 28 INFORMS INFORMS holds copyright to this article and distributed

More information

A framework for performance monitoring, load balancing, adaptive timeouts and quality of service in digital libraries

A framework for performance monitoring, load balancing, adaptive timeouts and quality of service in digital libraries Int J Digit Libr (2000) 3: 9 35 INTERNATIONAL JOURNAL ON Digital Libraries Springer-Verlag 2000 A fraework for perforance onitoring, load balancing, adaptive tieouts and quality of service in digital libraries

More information

Searching strategy for multi-target discovery in wireless networks

Searching strategy for multi-target discovery in wireless networks Searching strategy for ulti-target discovery in wireless networks Zhao Cheng, Wendi B. Heinzelan Departent of Electrical and Coputer Engineering University of Rochester Rochester, NY 467 (585) 75-{878,

More information

Resource Allocation in Wireless Networks with Multiple Relays

Resource Allocation in Wireless Networks with Multiple Relays Resource Allocation in Wireless Networks with Multiple Relays Kağan Bakanoğlu, Stefano Toasin, Elza Erkip Departent of Electrical and Coputer Engineering, Polytechnic Institute of NYU, Brooklyn, NY, 0

More information

Analyzing Spatiotemporal Characteristics of Education Network Traffic with Flexible Multiscale Entropy

Analyzing Spatiotemporal Characteristics of Education Network Traffic with Flexible Multiscale Entropy Vol. 9, No. 5 (2016), pp.303-312 http://dx.doi.org/10.14257/ijgdc.2016.9.5.26 Analyzing Spatioteporal Characteristics of Education Network Traffic with Flexible Multiscale Entropy Chen Yang, Renjie Zhou

More information

The Benefit of SMT in the Multi-Core Era: Flexibility towards Degrees of Thread-Level Parallelism

The Benefit of SMT in the Multi-Core Era: Flexibility towards Degrees of Thread-Level Parallelism The enefit of SMT in the Multi-Core Era: Flexibility towards Degrees of Thread-Level Parallelis Stijn Eyeran Lieven Eeckhout Ghent University, elgiu Stijn.Eyeran@elis.UGent.be, Lieven.Eeckhout@elis.UGent.be

More information

Applying Multiple Neural Networks on Large Scale Data

Applying Multiple Neural Networks on Large Scale Data 0 International Conference on Inforation and Electronics Engineering IPCSIT vol6 (0) (0) IACSIT Press, Singapore Applying Multiple Neural Networks on Large Scale Data Kritsanatt Boonkiatpong and Sukree

More information

CLOSED-LOOP SUPPLY CHAIN NETWORK OPTIMIZATION FOR HONG KONG CARTRIDGE RECYCLING INDUSTRY

CLOSED-LOOP SUPPLY CHAIN NETWORK OPTIMIZATION FOR HONG KONG CARTRIDGE RECYCLING INDUSTRY CLOSED-LOOP SUPPLY CHAIN NETWORK OPTIMIZATION FOR HONG KONG CARTRIDGE RECYCLING INDUSTRY Y. T. Chen Departent of Industrial and Systes Engineering Hong Kong Polytechnic University, Hong Kong yongtong.chen@connect.polyu.hk

More information

RECURSIVE DYNAMIC PROGRAMMING: HEURISTIC RULES, BOUNDING AND STATE SPACE REDUCTION. Henrik Kure

RECURSIVE DYNAMIC PROGRAMMING: HEURISTIC RULES, BOUNDING AND STATE SPACE REDUCTION. Henrik Kure RECURSIVE DYNAMIC PROGRAMMING: HEURISTIC RULES, BOUNDING AND STATE SPACE REDUCTION Henrik Kure Dina, Danish Inforatics Network In the Agricultural Sciences Royal Veterinary and Agricultural University

More information

An Integrated Approach for Monitoring Service Level Parameters of Software-Defined Networking

An Integrated Approach for Monitoring Service Level Parameters of Software-Defined Networking International Journal of Future Generation Counication and Networking Vol. 8, No. 6 (15), pp. 197-4 http://d.doi.org/1.1457/ijfgcn.15.8.6.19 An Integrated Approach for Monitoring Service Level Paraeters

More information

Energy Proportionality for Disk Storage Using Replication

Energy Proportionality for Disk Storage Using Replication Energy Proportionality for Disk Storage Using Replication Jinoh Ki and Doron Rote Lawrence Berkeley National Laboratory University of California, Berkeley, CA 94720 {jinohki,d rote}@lbl.gov Abstract Energy

More information

Preference-based Search and Multi-criteria Optimization

Preference-based Search and Multi-criteria Optimization Fro: AAAI-02 Proceedings. Copyright 2002, AAAI (www.aaai.org). All rights reserved. Preference-based Search and Multi-criteria Optiization Ulrich Junker ILOG 1681, route des Dolines F-06560 Valbonne ujunker@ilog.fr

More information

Media Adaptation Framework in Biofeedback System for Stroke Patient Rehabilitation

Media Adaptation Framework in Biofeedback System for Stroke Patient Rehabilitation Media Adaptation Fraework in Biofeedback Syste for Stroke Patient Rehabilitation Yinpeng Chen, Weiwei Xu, Hari Sundara, Thanassis Rikakis, Sheng-Min Liu Arts, Media and Engineering Progra Arizona State

More information

Machine Learning Applications in Grid Computing

Machine Learning Applications in Grid Computing Machine Learning Applications in Grid Coputing George Cybenko, Guofei Jiang and Daniel Bilar Thayer School of Engineering Dartouth College Hanover, NH 03755, USA gvc@dartouth.edu, guofei.jiang@dartouth.edu

More information

Impact of Processing Costs on Service Chain Placement in Network Functions Virtualization

Impact of Processing Costs on Service Chain Placement in Network Functions Virtualization Ipact of Processing Costs on Service Chain Placeent in Network Functions Virtualization Marco Savi, Massio Tornatore, Giacoo Verticale Dipartiento di Elettronica, Inforazione e Bioingegneria, Politecnico

More information

Use of extrapolation to forecast the working capital in the mechanical engineering companies

Use of extrapolation to forecast the working capital in the mechanical engineering companies ECONTECHMOD. AN INTERNATIONAL QUARTERLY JOURNAL 2014. Vol. 1. No. 1. 23 28 Use of extrapolation to forecast the working capital in the echanical engineering copanies A. Cherep, Y. Shvets Departent of finance

More information

Software Quality Characteristics Tested For Mobile Application Development

Software Quality Characteristics Tested For Mobile Application Development Thesis no: MGSE-2015-02 Software Quality Characteristics Tested For Mobile Application Developent Literature Review and Epirical Survey WALEED ANWAR Faculty of Coputing Blekinge Institute of Technology

More information

SOME APPLICATIONS OF FORECASTING Prof. Thomas B. Fomby Department of Economics Southern Methodist University May 2008

SOME APPLICATIONS OF FORECASTING Prof. Thomas B. Fomby Department of Economics Southern Methodist University May 2008 SOME APPLCATONS OF FORECASTNG Prof. Thoas B. Foby Departent of Econoics Southern Methodist University May 8 To deonstrate the usefulness of forecasting ethods this note discusses four applications of forecasting

More information

Evaluating Inventory Management Performance: a Preliminary Desk-Simulation Study Based on IOC Model

Evaluating Inventory Management Performance: a Preliminary Desk-Simulation Study Based on IOC Model Evaluating Inventory Manageent Perforance: a Preliinary Desk-Siulation Study Based on IOC Model Flora Bernardel, Roberto Panizzolo, and Davide Martinazzo Abstract The focus of this study is on preliinary

More information

An Approach to Combating Free-riding in Peer-to-Peer Networks

An Approach to Combating Free-riding in Peer-to-Peer Networks An Approach to Cobating Free-riding in Peer-to-Peer Networks Victor Ponce, Jie Wu, and Xiuqi Li Departent of Coputer Science and Engineering Florida Atlantic University Boca Raton, FL 33431 April 7, 2008

More information

The Application of Bandwidth Optimization Technique in SLA Negotiation Process

The Application of Bandwidth Optimization Technique in SLA Negotiation Process The Application of Bandwidth Optiization Technique in SLA egotiation Process Srecko Krile University of Dubrovnik Departent of Electrical Engineering and Coputing Cira Carica 4, 20000 Dubrovnik, Croatia

More information

Reliability Constrained Packet-sizing for Linear Multi-hop Wireless Networks

Reliability Constrained Packet-sizing for Linear Multi-hop Wireless Networks Reliability Constrained acket-sizing for inear Multi-hop Wireless Networks Ning Wen, and Randall A. Berry Departent of Electrical Engineering and Coputer Science Northwestern University, Evanston, Illinois

More information

SUPPORTING YOUR HIPAA COMPLIANCE EFFORTS

SUPPORTING YOUR HIPAA COMPLIANCE EFFORTS WHITE PAPER SUPPORTING YOUR HIPAA COMPLIANCE EFFORTS Quanti Solutions. Advancing HIM through Innovation HEALTHCARE SUPPORTING YOUR HIPAA COMPLIANCE EFFORTS Quanti Solutions. Advancing HIM through Innovation

More information

How To Balance Over Redundant Wireless Sensor Networks Based On Diffluent

How To Balance Over Redundant Wireless Sensor Networks Based On Diffluent Load balancing over redundant wireless sensor networks based on diffluent Abstract Xikui Gao Yan ai Yun Ju School of Control and Coputer Engineering North China Electric ower University 02206 China Received

More information

Data Set Generation for Rectangular Placement Problems

Data Set Generation for Rectangular Placement Problems Data Set Generation for Rectangular Placeent Probles Christine L. Valenzuela (Muford) Pearl Y. Wang School of Coputer Science & Inforatics Departent of Coputer Science MS 4A5 Cardiff University George

More information

Markovian inventory policy with application to the paper industry

Markovian inventory policy with application to the paper industry Coputers and Cheical Engineering 26 (2002) 1399 1413 www.elsevier.co/locate/copcheeng Markovian inventory policy with application to the paper industry K. Karen Yin a, *, Hu Liu a,1, Neil E. Johnson b,2

More information

Implementation of Active Queue Management in a Combined Input and Output Queued Switch

Implementation of Active Queue Management in a Combined Input and Output Queued Switch pleentation of Active Queue Manageent in a obined nput and Output Queued Switch Bartek Wydrowski and Moshe Zukeran AR Special Research entre for Ultra-Broadband nforation Networks, EEE Departent, The University

More information

Markov Models and Their Use for Calculations of Important Traffic Parameters of Contact Center

Markov Models and Their Use for Calculations of Important Traffic Parameters of Contact Center Markov Models and Their Use for Calculations of Iportant Traffic Paraeters of Contact Center ERIK CHROMY, JAN DIEZKA, MATEJ KAVACKY Institute of Telecounications Slovak University of Technology Bratislava

More information

Method of supply chain optimization in E-commerce

Method of supply chain optimization in E-commerce MPRA Munich Personal RePEc Archive Method of supply chain optiization in E-coerce Petr Suchánek and Robert Bucki Silesian University - School of Business Adinistration, The College of Inforatics and Manageent

More information

Efficient Key Management for Secure Group Communications with Bursty Behavior

Efficient Key Management for Secure Group Communications with Bursty Behavior Efficient Key Manageent for Secure Group Counications with Bursty Behavior Xukai Zou, Byrav Raaurthy Departent of Coputer Science and Engineering University of Nebraska-Lincoln Lincoln, NE68588, USA Eail:

More information

Image restoration for a rectangular poor-pixels detector

Image restoration for a rectangular poor-pixels detector Iage restoration for a rectangular poor-pixels detector Pengcheng Wen 1, Xiangjun Wang 1, Hong Wei 2 1 State Key Laboratory of Precision Measuring Technology and Instruents, Tianjin University, China 2

More information

REQUIREMENTS FOR A COMPUTER SCIENCE CURRICULUM EMPHASIZING INFORMATION TECHNOLOGY SUBJECT AREA: CURRICULUM ISSUES

REQUIREMENTS FOR A COMPUTER SCIENCE CURRICULUM EMPHASIZING INFORMATION TECHNOLOGY SUBJECT AREA: CURRICULUM ISSUES REQUIREMENTS FOR A COMPUTER SCIENCE CURRICULUM EMPHASIZING INFORMATION TECHNOLOGY SUBJECT AREA: CURRICULUM ISSUES Charles Reynolds Christopher Fox reynolds @cs.ju.edu fox@cs.ju.edu Departent of Coputer

More information

Modeling Parallel Applications Performance on Heterogeneous Systems

Modeling Parallel Applications Performance on Heterogeneous Systems Modeling Parallel Applications Perforance on Heterogeneous Systes Jaeela Al-Jaroodi, Nader Mohaed, Hong Jiang and David Swanson Departent of Coputer Science and Engineering University of Nebraska Lincoln

More information

Data Streaming Algorithms for Estimating Entropy of Network Traffic

Data Streaming Algorithms for Estimating Entropy of Network Traffic Data Streaing Algoriths for Estiating Entropy of Network Traffic Ashwin Lall University of Rochester Vyas Sekar Carnegie Mellon University Mitsunori Ogihara University of Rochester Jun (Ji) Xu Georgia

More information

Protecting Small Keys in Authentication Protocols for Wireless Sensor Networks

Protecting Small Keys in Authentication Protocols for Wireless Sensor Networks Protecting Sall Keys in Authentication Protocols for Wireless Sensor Networks Kalvinder Singh Australia Developent Laboratory, IBM and School of Inforation and Counication Technology, Griffith University

More information

Evaluating the Effectiveness of Task Overlapping as a Risk Response Strategy in Engineering Projects

Evaluating the Effectiveness of Task Overlapping as a Risk Response Strategy in Engineering Projects Evaluating the Effectiveness of Task Overlapping as a Risk Response Strategy in Engineering Projects Lucas Grèze Robert Pellerin Nathalie Perrier Patrice Leclaire February 2011 CIRRELT-2011-11 Bureaux

More information

PREDICTION OF POSSIBLE CONGESTIONS IN SLA CREATION PROCESS

PREDICTION OF POSSIBLE CONGESTIONS IN SLA CREATION PROCESS PREDICTIO OF POSSIBLE COGESTIOS I SLA CREATIO PROCESS Srećko Krile University of Dubrovnik Departent of Electrical Engineering and Coputing Cira Carica 4, 20000 Dubrovnik, Croatia Tel +385 20 445-739,

More information

Workflow Management in Cloud Computing

Workflow Management in Cloud Computing Workflow Manageent in Cloud Coputing Monika Bharti M.E. student Coputer Science and Engineering Departent Thapar University, Patiala Anju Bala Assistant Professor Coputer Science and Engineering Departent

More information

Managing Complex Network Operation with Predictive Analytics

Managing Complex Network Operation with Predictive Analytics Managing Coplex Network Operation with Predictive Analytics Zhenyu Huang, Pak Chung Wong, Patrick Mackey, Yousu Chen, Jian Ma, Kevin Schneider, and Frank L. Greitzer Pacific Northwest National Laboratory

More information

Considerations on Distributed Load Balancing for Fully Heterogeneous Machines: Two Particular Cases

Considerations on Distributed Load Balancing for Fully Heterogeneous Machines: Two Particular Cases Considerations on Distributed Load Balancing for Fully Heterogeneous Machines: Two Particular Cases Nathanaël Cheriere Departent of Coputer Science ENS Rennes Rennes, France nathanael.cheriere@ens-rennes.fr

More information

Approximately-Perfect Hashing: Improving Network Throughput through Efficient Off-chip Routing Table Lookup

Approximately-Perfect Hashing: Improving Network Throughput through Efficient Off-chip Routing Table Lookup Approxiately-Perfect ing: Iproving Network Throughput through Efficient Off-chip Routing Table Lookup Zhuo Huang, Jih-Kwon Peir, Shigang Chen Departent of Coputer & Inforation Science & Engineering, University

More information

Fuzzy Sets in HR Management

Fuzzy Sets in HR Management Acta Polytechnica Hungarica Vol. 8, No. 3, 2011 Fuzzy Sets in HR Manageent Blanka Zeková AXIOM SW, s.r.o., 760 01 Zlín, Czech Republic blanka.zekova@sezna.cz Jana Talašová Faculty of Science, Palacký Univerzity,

More information

Information Processing Letters

Information Processing Letters Inforation Processing Letters 111 2011) 178 183 Contents lists available at ScienceDirect Inforation Processing Letters www.elsevier.co/locate/ipl Offline file assignents for online load balancing Paul

More information

Design of Model Reference Self Tuning Mechanism for PID like Fuzzy Controller

Design of Model Reference Self Tuning Mechanism for PID like Fuzzy Controller Research Article International Journal of Current Engineering and Technology EISSN 77 46, PISSN 347 56 4 INPRESSCO, All Rights Reserved Available at http://inpressco.co/category/ijcet Design of Model Reference

More information

Partitioned Elias-Fano Indexes

Partitioned Elias-Fano Indexes Partitioned Elias-ano Indexes Giuseppe Ottaviano ISTI-CNR, Pisa giuseppe.ottaviano@isti.cnr.it Rossano Venturini Dept. of Coputer Science, University of Pisa rossano@di.unipi.it ABSTRACT The Elias-ano

More information

The AGA Evaluating Model of Customer Loyalty Based on E-commerce Environment

The AGA Evaluating Model of Customer Loyalty Based on E-commerce Environment 6 JOURNAL OF SOFTWARE, VOL. 4, NO. 3, MAY 009 The AGA Evaluating Model of Custoer Loyalty Based on E-coerce Environent Shaoei Yang Econoics and Manageent Departent, North China Electric Power University,

More information

Load Control for Overloaded MPLS/DiffServ Networks during SLA Negotiation

Load Control for Overloaded MPLS/DiffServ Networks during SLA Negotiation Int J Counications, Network and Syste Sciences, 29, 5, 422-432 doi:14236/ijcns292547 Published Online August 29 (http://wwwscirporg/journal/ijcns/) Load Control for Overloaded MPLS/DiffServ Networks during

More information

INTEGRATED ENVIRONMENT FOR STORING AND HANDLING INFORMATION IN TASKS OF INDUCTIVE MODELLING FOR BUSINESS INTELLIGENCE SYSTEMS

INTEGRATED ENVIRONMENT FOR STORING AND HANDLING INFORMATION IN TASKS OF INDUCTIVE MODELLING FOR BUSINESS INTELLIGENCE SYSTEMS Artificial Intelligence Methods and Techniques for Business and Engineering Applications 210 INTEGRATED ENVIRONMENT FOR STORING AND HANDLING INFORMATION IN TASKS OF INDUCTIVE MODELLING FOR BUSINESS INTELLIGENCE

More information

A Study on the Chain Restaurants Dynamic Negotiation Games of the Optimization of Joint Procurement of Food Materials

A Study on the Chain Restaurants Dynamic Negotiation Games of the Optimization of Joint Procurement of Food Materials International Journal of Coputer Science & Inforation Technology (IJCSIT) Vol 6, No 1, February 2014 A Study on the Chain estaurants Dynaic Negotiation aes of the Optiization of Joint Procureent of Food

More information

Modeling Nurse Scheduling Problem Using 0-1 Goal Programming: A Case Study Of Tafo Government Hospital, Kumasi-Ghana

Modeling Nurse Scheduling Problem Using 0-1 Goal Programming: A Case Study Of Tafo Government Hospital, Kumasi-Ghana Modeling Nurse Scheduling Proble Using 0-1 Goal Prograing: A Case Study Of Tafo Governent Hospital, Kuasi-Ghana Wallace Agyei, Willia Obeng-Denteh, Eanuel A. Andaa Abstract: The proble of scheduling nurses

More information

An Improved Decision-making Model of Human Resource Outsourcing Based on Internet Collaboration

An Improved Decision-making Model of Human Resource Outsourcing Based on Internet Collaboration International Journal of Hybrid Inforation Technology, pp. 339-350 http://dx.doi.org/10.14257/hit.2016.9.4.28 An Iproved Decision-aking Model of Huan Resource Outsourcing Based on Internet Collaboration

More information

Optimal Resource-Constraint Project Scheduling with Overlapping Modes

Optimal Resource-Constraint Project Scheduling with Overlapping Modes Optial Resource-Constraint Proect Scheduling with Overlapping Modes François Berthaut Lucas Grèze Robert Pellerin Nathalie Perrier Adnène Hai February 20 CIRRELT-20-09 Bureaux de Montréal : Bureaux de

More information

Research Article Performance Evaluation of Human Resource Outsourcing in Food Processing Enterprises

Research Article Performance Evaluation of Human Resource Outsourcing in Food Processing Enterprises Advance Journal of Food Science and Technology 9(2): 964-969, 205 ISSN: 2042-4868; e-issn: 2042-4876 205 Maxwell Scientific Publication Corp. Subitted: August 0, 205 Accepted: Septeber 3, 205 Published:

More information

Airline Yield Management with Overbooking, Cancellations, and No-Shows JANAKIRAM SUBRAMANIAN

Airline Yield Management with Overbooking, Cancellations, and No-Shows JANAKIRAM SUBRAMANIAN Airline Yield Manageent with Overbooking, Cancellations, and No-Shows JANAKIRAM SUBRAMANIAN Integral Developent Corporation, 301 University Avenue, Suite 200, Palo Alto, California 94301 SHALER STIDHAM

More information

Generating Certification Authority Authenticated Public Keys in Ad Hoc Networks

Generating Certification Authority Authenticated Public Keys in Ad Hoc Networks SECURITY AND COMMUNICATION NETWORKS Published online in Wiley InterScience (www.interscience.wiley.co). Generating Certification Authority Authenticated Public Keys in Ad Hoc Networks G. Kounga 1, C. J.

More information

An Optimal Task Allocation Model for System Cost Analysis in Heterogeneous Distributed Computing Systems: A Heuristic Approach

An Optimal Task Allocation Model for System Cost Analysis in Heterogeneous Distributed Computing Systems: A Heuristic Approach An Optial Tas Allocation Model for Syste Cost Analysis in Heterogeneous Distributed Coputing Systes: A Heuristic Approach P. K. Yadav Central Building Research Institute, Rooree- 247667, Uttarahand (INDIA)

More information

Dynamic right-sizing for power-proportional data centers Extended version

Dynamic right-sizing for power-proportional data centers Extended version 1 Dynaic right-sizing for power-proportional data centers Extended version Minghong Lin, Ada Wieran, Lachlan L. H. Andrew and Eno Thereska Abstract Power consuption iposes a significant cost for data centers

More information

A Fast Algorithm for Online Placement and Reorganization of Replicated Data

A Fast Algorithm for Online Placement and Reorganization of Replicated Data A Fast Algorith for Online Placeent and Reorganization of Replicated Data R. J. Honicky Storage Systes Research Center University of California, Santa Cruz Ethan L. Miller Storage Systes Research Center

More information

CRM FACTORS ASSESSMENT USING ANALYTIC HIERARCHY PROCESS

CRM FACTORS ASSESSMENT USING ANALYTIC HIERARCHY PROCESS 641 CRM FACTORS ASSESSMENT USING ANALYTIC HIERARCHY PROCESS Marketa Zajarosova 1* *Ph.D. VSB - Technical University of Ostrava, THE CZECH REPUBLIC arketa.zajarosova@vsb.cz Abstract Custoer relationship

More information

Sensors as a Service Oriented Architecture: Middleware for Sensor Networks

Sensors as a Service Oriented Architecture: Middleware for Sensor Networks Sensors as a Service Oriented Architecture: Middleware for Sensor Networks John Ibbotson, Christopher Gibson, Joel Wright, Peter Waggett, IBM U.K Ltd, Petros Zerfos, IBM Research, Boleslaw K. Szyanski,

More information

Halloween Costume Ideas for the Wii Game

Halloween Costume Ideas for the Wii Game Algorithica 2001) 30: 101 139 DOI: 101007/s00453-001-0003-0 Algorithica 2001 Springer-Verlag New York Inc Optial Search and One-Way Trading Online Algoriths R El-Yaniv, 1 A Fiat, 2 R M Karp, 3 and G Turpin

More information

Equivalent Tapped Delay Line Channel Responses with Reduced Taps

Equivalent Tapped Delay Line Channel Responses with Reduced Taps Equivalent Tapped Delay Line Channel Responses with Reduced Taps Shweta Sagari, Wade Trappe, Larry Greenstein {shsagari, trappe, ljg}@winlab.rutgers.edu WINLAB, Rutgers University, North Brunswick, NJ

More information

A short-term, pattern-based model for water-demand forecasting

A short-term, pattern-based model for water-demand forecasting 39 Q IWA Publishing 2007 Journal of Hydroinforatics 09.1 2007 A short-ter, pattern-based odel for water-deand forecasting Stefano Alvisi, Marco Franchini and Alberto Marinelli ABSTRACT The short-ter, deand-forecasting

More information

Reconnect 04 Solving Integer Programs with Branch and Bound (and Branch and Cut)

Reconnect 04 Solving Integer Programs with Branch and Bound (and Branch and Cut) Sandia is a ultiprogra laboratory operated by Sandia Corporation, a Lockheed Martin Copany, Reconnect 04 Solving Integer Progras with Branch and Bound (and Branch and Cut) Cynthia Phillips (Sandia National

More information

Introduction to the Microsoft Sync Framework. Michael Clark Development Manager Microsoft

Introduction to the Microsoft Sync Framework. Michael Clark Development Manager Microsoft Introduction to the Michael Clark Developent Manager Microsoft Agenda Why Is Sync both Interesting and Hard Sync Fraework Overview Using the Sync Fraework Future Directions Suary Why Is Sync Iportant Coputing

More information

Red Hat Enterprise Linux: Creating a Scalable Open Source Storage Infrastructure

Red Hat Enterprise Linux: Creating a Scalable Open Source Storage Infrastructure Red Hat Enterprise Linux: Creating a Scalable Open Source Storage Infrastructure By Alan Radding and Nick Carr Abstract This paper discusses the issues related to storage design and anageent when an IT

More information

Online Algorithm for Servers Consolidation in Cloud Data Centers

Online Algorithm for Servers Consolidation in Cloud Data Centers 1 Online Algorithm for Servers Consolidation in Cloud Data Centers Makhlouf Hadji, Paul Labrogere Technological Research Institute - IRT SystemX 8, Avenue de la Vauve, 91120 Palaiseau, France. Abstract

More information

A CHAOS MODEL OF SUBHARMONIC OSCILLATIONS IN CURRENT MODE PWM BOOST CONVERTERS

A CHAOS MODEL OF SUBHARMONIC OSCILLATIONS IN CURRENT MODE PWM BOOST CONVERTERS A CHAOS MODEL OF SUBHARMONIC OSCILLATIONS IN CURRENT MODE PWM BOOST CONVERTERS Isaac Zafrany and Sa BenYaakov Departent of Electrical and Coputer Engineering BenGurion University of the Negev P. O. Box

More information

Audio Engineering Society. Convention Paper. Presented at the 119th Convention 2005 October 7 10 New York, New York USA

Audio Engineering Society. Convention Paper. Presented at the 119th Convention 2005 October 7 10 New York, New York USA Audio Engineering Society Convention Paper Presented at the 119th Convention 2005 October 7 10 New York, New York USA This convention paper has been reproduced fro the authors advance anuscript, without

More information

The Virtual Spring Mass System

The Virtual Spring Mass System The Virtual Spring Mass Syste J. S. Freudenberg EECS 6 Ebedded Control Systes Huan Coputer Interaction A force feedbac syste, such as the haptic heel used in the EECS 6 lab, is capable of exhibiting a

More information

ON SELF-ROUTING IN CLOS CONNECTION NETWORKS. BARRY G. DOUGLASS Electrical Engineering Department Texas A&M University College Station, TX 77843-3128

ON SELF-ROUTING IN CLOS CONNECTION NETWORKS. BARRY G. DOUGLASS Electrical Engineering Department Texas A&M University College Station, TX 77843-3128 ON SELF-ROUTING IN CLOS CONNECTION NETWORKS BARRY G. DOUGLASS Electrical Engineering Departent Texas A&M University College Station, TX 778-8 A. YAVUZ ORUÇ Electrical Engineering Departent and Institute

More information

A Multi-Core Pipelined Architecture for Parallel Computing

A Multi-Core Pipelined Architecture for Parallel Computing Parallel & Cloud Coputing PCC Vol, Iss A Multi-Core Pipelined Architecture for Parallel Coputing Duoduo Liao *1, Sion Y Berkovich Coputing for Geospatial Research Institute Departent of Coputer Science,

More information

Heterogeneous Workload Consolidation for Efficient Management of Data Centers in Cloud Computing

Heterogeneous Workload Consolidation for Efficient Management of Data Centers in Cloud Computing Heterogeneous Workload Consolidation for Efficient Management of Data Centers in Cloud Computing Deep Mann ME (Software Engineering) Computer Science and Engineering Department Thapar University Patiala-147004

More information

Network delay-aware load balancing in selfish and cooperative distributed systems

Network delay-aware load balancing in selfish and cooperative distributed systems Network delay-aware load balancing in selfish and cooperative distributed systes Piotr Skowron Faculty of Matheatics, Inforatics and Mechanics University of Warsaw Eail: p.skowron@iuw.edu.pl Krzysztof

More information

The individual neurons are complicated. They have a myriad of parts, subsystems and control mechanisms. They convey information via a host of

The individual neurons are complicated. They have a myriad of parts, subsystems and control mechanisms. They convey information via a host of CHAPTER 4 ARTIFICIAL NEURAL NETWORKS 4. INTRODUCTION Artificial Neural Networks (ANNs) are relatively crude electronic odels based on the neural structure of the brain. The brain learns fro experience.

More information

Position Auctions and Non-uniform Conversion Rates

Position Auctions and Non-uniform Conversion Rates Position Auctions and Non-unifor Conversion Rates Liad Blurosen Microsoft Research Mountain View, CA 944 liadbl@icrosoft.co Jason D. Hartline Shuzhen Nong Electrical Engineering and Microsoft AdCenter

More information

Performance Evaluation of Machine Learning Techniques using Software Cost Drivers

Performance Evaluation of Machine Learning Techniques using Software Cost Drivers Perforance Evaluation of Machine Learning Techniques using Software Cost Drivers Manas Gaur Departent of Coputer Engineering, Delhi Technological University Delhi, India ABSTRACT There is a treendous rise

More information

Models and Algorithms for Stochastic Online Scheduling 1

Models and Algorithms for Stochastic Online Scheduling 1 Models and Algoriths for Stochastic Online Scheduling 1 Nicole Megow Technische Universität Berlin, Institut für Matheatik, Strasse des 17. Juni 136, 10623 Berlin, Gerany. eail: negow@ath.tu-berlin.de

More information

AN ALGORITHM FOR REDUCING THE DIMENSION AND SIZE OF A SAMPLE FOR DATA EXPLORATION PROCEDURES

AN ALGORITHM FOR REDUCING THE DIMENSION AND SIZE OF A SAMPLE FOR DATA EXPLORATION PROCEDURES Int. J. Appl. Math. Coput. Sci., 2014, Vol. 24, No. 1, 133 149 DOI: 10.2478/acs-2014-0011 AN ALGORITHM FOR REDUCING THE DIMENSION AND SIZE OF A SAMPLE FOR DATA EXPLORATION PROCEDURES PIOTR KULCZYCKI,,

More information

arxiv:0805.1434v1 [math.pr] 9 May 2008

arxiv:0805.1434v1 [math.pr] 9 May 2008 Degree-distribution stability of scale-free networs Zhenting Hou, Xiangxing Kong, Dinghua Shi,2, and Guanrong Chen 3 School of Matheatics, Central South University, Changsha 40083, China 2 Departent of

More information

Factored Models for Probabilistic Modal Logic

Factored Models for Probabilistic Modal Logic Proceedings of the Twenty-Third AAAI Conference on Artificial Intelligence (2008 Factored Models for Probabilistic Modal Logic Afsaneh Shirazi and Eyal Air Coputer Science Departent, University of Illinois

More information

Local Area Network Management

Local Area Network Management Technology Guidelines for School Coputer-based Technologies Local Area Network Manageent Local Area Network Manageent Introduction This docuent discusses the tasks associated with anageent of Local Area

More information

AutoHelp. An 'Intelligent' Case-Based Help Desk Providing. Web-Based Support for EOSDIS Customers. A Concept and Proof-of-Concept Implementation

AutoHelp. An 'Intelligent' Case-Based Help Desk Providing. Web-Based Support for EOSDIS Customers. A Concept and Proof-of-Concept Implementation //j yd xd/_ ' Year One Report ":,/_i',:?,2... i" _.,.j- _,._".;-/._. ","/ AutoHelp An 'Intelligent' Case-Based Help Desk Providing Web-Based Support for EOSDIS Custoers A Concept and Proof-of-Concept Ipleentation

More information

( C) CLASS 10. TEMPERATURE AND ATOMS

( C) CLASS 10. TEMPERATURE AND ATOMS CLASS 10. EMPERAURE AND AOMS 10.1. INRODUCION Boyle s understanding of the pressure-volue relationship for gases occurred in the late 1600 s. he relationships between volue and teperature, and between

More information

ADJUSTING FOR QUALITY CHANGE

ADJUSTING FOR QUALITY CHANGE ADJUSTING FOR QUALITY CHANGE 7 Introduction 7.1 The easureent of changes in the level of consuer prices is coplicated by the appearance and disappearance of new and old goods and services, as well as changes

More information

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, ACCEPTED FOR PUBLICATION 1. Secure Wireless Multicast for Delay-Sensitive Data via Network Coding

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, ACCEPTED FOR PUBLICATION 1. Secure Wireless Multicast for Delay-Sensitive Data via Network Coding IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, ACCEPTED FOR PUBLICATION 1 Secure Wireless Multicast for Delay-Sensitive Data via Network Coding Tuan T. Tran, Meber, IEEE, Hongxiang Li, Senior Meber, IEEE,

More information

Stochastic Online Scheduling on Parallel Machines

Stochastic Online Scheduling on Parallel Machines Stochastic Online Scheduling on Parallel Machines Nicole Megow 1, Marc Uetz 2, and Tark Vredeveld 3 1 Technische Universit at Berlin, Institut f ur Matheatik, Strasse des 17. Juni 136, 10623 Berlin, Gerany

More information

Endogenous Market Structure and the Cooperative Firm

Endogenous Market Structure and the Cooperative Firm Endogenous Market Structure and the Cooperative Fir Brent Hueth and GianCarlo Moschini Working Paper 14-WP 547 May 2014 Center for Agricultural and Rural Developent Iowa State University Aes, Iowa 50011-1070

More information