Optimal Data Distribution for Heterogeneous Parallel Storage Servers Streaming Media Files

Size: px
Start display at page:

Download "Optimal Data Distribution for Heterogeneous Parallel Storage Servers Streaming Media Files"

Transcription

1 Optimal Data Distribution or Heterogeneous Parallel Storage Servers Streaming Meia Files Christian Ortol Department o Computer Science, University o Freiburg, Georges-Koehler-Allee 5, 79 Freiburg, Germany ortol@inormatik.uni-reiburg.e Christian Schinelhauer Department o Computer Science, University o Freiburg, Georges-Koehler-Allee 5, 79 Freiburg, Germany schinel@inormatik.uni-reiburg.e Abstract We consier the problem o istributing meia iles or streaming on a istribute storage network, where servers have heterogeneous capacities an banwiths. Regaring networking the servers banwiths are the bottlenecks or streaming. We present an algorithm that computes an assignment o n iles to m servers or istributing meia iles such that the streaming spee requirements an capacity constraints are kept. As an aitional eature this assignment algorithm works on, i.e. it can assign each ile without iles to be store later on. Our algorithm computes the ata assignment in O(nm+m log m) outperorming ar program solvers. I. INTRODUCTION Streaming meia iles has become a stanar application or PCs an smart TVs in moern homes. Currently, this requires a eicate streaming server powerul enough to provie the banwith or high einition mulia iles an at the same have enough storage capacity to contain a large meia library. Distribute ile systems coul rener such servers unnecessary by accumulating the require resources rom multiple potentially less powerul evices like noes in a boy area network (BAN), which may be too slow or equippe with too little memory or inepenently serving the meia iles alone. Distribute ile systems have the avantage that they can be built to scale or many large iles or access by many clients in parallel or minimize parallel rea o a ile. None o these optimizations solve the problem pose by the combination o banwith requirement with a restricte storage on evices with heterogeneous banwith capabilities. Here, we solve the istribution o ile ragments across a heterogeneous network o ile servers. This requires a partition o a meia ile, an the assignment o the parts to meia servers, which on eman serve their parts with a ixe known bit rate. The necessity o meia iles istribution comes rom resource restricte storage servers with limite storage capacity an storage access rates. Our objective is to partition an assign the meia iles to the servers, either o or on. On means that the iles arrive in a ranom orer an nee to be assigne once or all. In the o case, we know the sizes an streaming rates o all iles to come an want to compute an optimal strategy. In a homogenous setting it is obvious that a ile can just be partitionento equal parts an place on all server with same size an spee. But with servers varying in storage capacity an banwith, an equal istribution may clog up high banwith or low storage servers very quickly. We show in this paper how to achieve an optimal istribution on. Thereore we can solve this ile istribution problem without incurring any cost or reistributing iles later on. This perect istribution o ata in an inocommunication system can be use as a tool to eiciently hanle ata or representation-briging communication an meiastreaming, e.g. sensory ata can be istribute an storen a BAN to later be playe back to its user. Relate Work.: Since the introuction o reunant arrays o inepenent iscs in the seminal paper by Patterson et al. [7] istribute storage has evolve. While RAID systems o not know heterogeneous isc sizes, network connections, hotspot iles, scalability an optimizing energy consumption, moern istribute storage systems consier such actors. The Google File System (GFS) [] was esigne to satisy the nee or a highly scalable an high throughput system or Google s web search service. It uses a large number o commoity computers (chunk servers) combine with a single controlling server in a homogenous network setting. Files are split into chunks an istribute to the chunk servers. For ile retrieval a client queries the location o the chunks rom the master server an ownloas the chunk irectly rom the chunk server. Parallelization comes into play when multiple clients access ierent chunk servers an hot spot chunks are copie to multiple chunk servers. Dynamo [] was evelope or Amazon an ocuses on parallel write operations an high availability. Instea o a central master server, it uses consistent hashing [] to istribute ata onto the servers. By utilizing virtual copies o noes in the consistent hashing it can hanle servers with heterogeneous storage capacities. Like GFS it oes not hanle heterogeneous banwiths as all servers are assume to be in the same local network. Various other istribute storage systems utilize peer-topeer networks [], [5], []. These works ocus on ierent challenges like the reputation o peers, the reliability o store iles with oten isconnecte peers or proviing byzantine ault tolerance in the network. The assumption o an homogenous unerlying network inrastructure restricts the applicability o these systems. Storage inrastructures ten to evolve over an become heterogeneous, which stanar solutions o not cover. Likewise any Internet base service will inevitably ace a large variety o ierent connections.

2 The Distribute Parallel File System (DPFS) by Shen an Chouhary [] was built to hanle such an heterogeneous networking inrastructure. Files to be store are broken up into chunks an istribute among the network o chunk servers. DPFS ocuses on how the chunks shoul be istribute or optimal perormance uner expecte access pattern, but they also accommoate or ierent banwiths by istributing chunks proportionally to a given perormance actor among servers using a simple greey algorithm. While consistent hashing can be aapte with the help o virtual servers to heterogeneous settings, Distribute Heterogeneous Hash Tables (DHHT) [8] are alreay esigne or heterogeneity. Servers can be weighte accoring to capacity or connection banwith an receive a larger number o chunks. The DHHT was later utilizen a istribute storage system in [9]. There an optimal solution or the average parallel transer o a ocument is shown. In the work by Langner et al. [6] the uploa o chunks in a setting with asymmetric connections is consiere. They show a solution that also minimizes the parallel or uploa an ownloa. The moel presenten this paper is base on the master thesis o Schott []. He presents a similar, but less eicient algorithm. II. THE PROBLEM SETTING AND CONTRIBUTION We consier a ixe set o m servers,..., s m with storage capacities c,..., c m measuren bits. Let b i enote the banwith o a server, i.e. the number o bits a storage server can stream per secon. Now, the meia iles,..., n are ae to the storage servers, where j enotes the size o ile j an j the necessary streaming ata rate o this ile. We assume that the storage server is the bottleneck. So, the or ownloaing the ile i is etermine by the slowest part. Now the ile is assigne to the servers inicate by the assignment matrix A = (a i,j ) i [n],j [m with a i,j inicates the number o bits o ile i store on storage server s j. Clearly, all bits o the ile nee to be istribute on all storage servers. m a i,j = i or all i [n]. () j= O course, the capacity o the server cannot be exceee. n a i,j c j or all j [m]. () i= We assume that the banwith o the storage server is the bottleneck, while the network oes not pose urther constraints. From the minimum ata rate o a ile i it is clear that the maximum it may take to retrieve a ile is i /. For each server s j the minimum to sents parts o i is a i,j /. For continuous streaming we nee a i,j / i / or each assignment. This leas to the main constraint o the parallel heterogeneous streaming problem: O course, we also nee to consier where to place each bit o ata. For this, the ile nees to be partitionento blocks o aequate size resulting rom a ar selection o the ata. There is a straight-orwar solution or this problem an thereore we o not consier it in this paper. Here, we concentrate only on the assignment problem. Deinition The parallel, heterogeneous o streaming assignment problem is, given servers,..., s m an all iles,..., n, compute an assignment A which satisies constraints (), () an (). This is clearly a ar program. We solve this problem in O(nm + m log m), which is aster than any existing ar program solver. For the on streaming assignment problem servers,..., s m are given at the beginning. The iles,..., n are given sequentially, such that the assignment a i,j or ile i must be compute beore iles i+,..., n are given. The task is to compute an assignment which satisies constraints (), () an () i possible. We call an on streaming assignment perect, i it can assign the same number o iles as an o assignment. Our algorithm is an on algorithm, an thus perect. III. AN EFFICIENT OFFLINE AND ONLINE SOLUTION The key to our solution is the notion o sustainability. Deinition The sustainability σ j o a storage server s j is eine as the it nees to rea out all ata, i.e. σ j := c j /. Servers with equal sustainability can be split an joine without aecting the solvability o the problem. Lemma For any set o iles the assignment problem onto servers S = {,..., s m }, where σ m = σ m can be solve an only i the assignment problem to servers S = {,..., s m, s } can be solve, where the capacity o s is c m + c m an the banwith o s is b = b m + b m. Proo: Assume that (a i,j ) i [n],j [m] is an assignment o the iles to S. Then the assignment (a i,j ) i [n],j [m] to S is vali where { a ai,j, j [m ] i,j = a i,m + a i,m, j = m Clearly, constraints () an () hol. For constraint () we observe a i,m + a i,m i b m + b m = i b. Now assume that an assignment (a i,j ) i [n],j [m] or S is given, then an assignment or S can be compute as ollows, j [m ] a i,j = a i,j a i,m b m b m + b m, j = m a i,j i or all i [n], j [m]. () a i,m b m b m + b m, j = m

3 Again constraints () an () are straight-orwar. For constraint () we have b m a i,m = a i,m b m + b m an analogously or a i,m. i b m The irst step to solve the assignment problem is to sort all storage servers with respect to their sustainability σ i, such that σ σ... or storage servers,..., s m. Let t i = i / enote the play o ile i. In real-worl applications iles are continuously ae or remove rom storage server systems. Then, it is costly to reistribute an recompute all assignments when a new ile arrives. So, we consier the case where assignments are mae or each ile separately without the knowlege o uture iles. Hence, rearranging the set o iles, e.g. accoring to their play is not allowe. One might think that this reuces the possible solution space. However, we present with Algorithm an on algorithm which also computes vali assignments or the o problem. So, this algorithm reveals a structural property o the assignment problem. The algorithm sorts all servers accoring to their sustainability. Then it computes or each ile i an assignment in the or-loop between s 7 an 5. The basic iea is a that moves over the storage servers epicten Fig.. The vertical moves continuously rom the right to the let. When it moves over the rectangle representing a server the area to the right o the inicates the assigne amount o ata. However, i it has reache an area o size i /, which correspons to istance t i ater touching a rectangle or the irst, it will stop the assignment to this server because o constraint (). The stops, i the has collecte area o size i or it reaches the vertical axis. In the irst case assignment o the ile can be compute, in the secon case there is no assignment. An eicient implementation o such a technique uses so-calle events. These events are the beginning o rectangles at σ i,j+, the constraint () at σ i,j+ t i, or the halt o the when the complete ile is assigne to the servers in 7. The starts rom the maximum value σ i, an chooses the next event by maximizing over these three cases. Then, the next rectangle will be ae to the assignments in 6. The currently active servers in the while loop are s j,..., s j. Each event nees a special treatment or the next roun, i.e. removing a storage server by increasing j, aing a storage server by increasing j. The stops when all storage is assigne or it runs out o servers j > m. The ollowing notations are useul or analyzing the algorithm. Deine the resiual storage c i,j ater inserting i iles as ollows. c,j := c j or all j [m] c i,j := c i,j a i,j or all i [n], j [m] Algorithm Sweep Line Algorithm or Streaming Assignment Input: storage servers,..., s m with capacity c,..., c m an banwith,..., b m an iles,..., n with ata rates,..., n : Sort storage servers such that σ σ... σ m : or j o m o : σ,j c j / : c,j c j 5: en or 6: σ,m+ 7: or i o n o 8: or j o m o 9: a i,j : en or : t i i / : j : j /* R R {i} */ : s 5: t last σ i, 6: while s < i an j m an t new o 7: t new σ i,j t i 8: i j < m then 9: t new max{t new, σ i,j+} : en : i j j hen{ } : t new max t new, t last i s j j=j : en : i t new then 5: or j j o j o 6: a i,j a i,j + (t last t new ) 7: s s + (t last t new ) 8: en or 9: t last t new : i t new = σ i,j t i then : j j + /* R j R j \ {i} : F j F j {i} */ : else i t new = σ i,j+hen : j j + /* R j R j {i} */ 5: en 6: en 7: or j o m o 8: σ i,j σ i,j a i,j / 9: en or : σ i,m+ : en while : i j > m or t new < then : return File cannot be assigne : en 5: en or 6: return (a i,j ) j [m]

4 The resiual sustainability o the server j ater inserting i iles is enote by σ i,j. σ i,j := c i,j or all i [n], j [m] The assignment o a server measuren play is enote by z i,j. z i,j := a i,j or all i [n], j [m] Note that constraint () translates into the conition z i,j t i or all i, j. In the commentary section the assignments are classiie by the sets R k = {i : < z i,k < t i } or all k [m] F k = {i : z i,k = t i } or all k [m] where F k (ull play) enotes the set o ile inices where the maximum block size is assigne to server s k accoring constraint (). R k (rest o the ile) enotes the set o ile inices with other non-zero assignments o blocks to server s k. I these sets are given, then the assignment can be presente in close orm, see Algorithm. Algorithm Alternative presentation o Algorithm Input: storage servers,..., s m with capacity c,,..., c,m sorte accoring σ j = c j / in escening orer banwith,..., b m an iles,..., n with ata rates,..., n sets R,..., R m, F,..., F m rom Algorithm : or i = o n o : or all j : i F j o : a i,j i : en or j 5: t :i F j t i + j :i R c j i,j i j :i R j 6: or all j : i R j o 7: a i,j c i,j t 8: en or 9: or j o m o : c i,j c i,j a i,j : en or : en or Lemma I Algorithm computes an assignment, then Algorithm computes the same assignment. Proo: I i F j, then z i,j = t i an thereore a i,j / = i /. I i R j, then the rest o the ile is istribute onto the servers j,..., j. For this the position t o the is compute which satisies (σ i,j t ) + t i = i j :i R j j :i F j The assigne storage rom a i,j is then given by (σ i,j t ). We prove the correctness o the algorithm by an inuction over the set o servers. Lemma Algorithm is correct or a single server. Proo: For one server Algorithm has only two kin o events or the : checking or constraint () an checking or constraint (). So, the algorithm ecies correctly, whether the iles can be assigne. Theorem The On Algorithm computes a vali assignment or every set o iles,..., n i it is possible. Thereore, it provies a perect solution to the on problem in O(nm + m log m). Proo: The sorting o all servers requires O(m log m). The sums j j=j usen can be precompute when the interval [j, j ] is change with amortize constant costs. For the run- o the while loop observe that in each roun either j or j is incremente or the loop terminates. Thereore, the inner statements o the loops are perorme m + imes. Thus, assigning each o the n ile takes O(m) steps. We will now prove that or any set o iles,..., n that Algorithm ins a vali assignment i possible. For this we take a closer look when a part o a ile is assigne to a server maximizing constraint (), i.e. i F k, or when it is below but not zero, i.e. i R k. I i F k+, then we have also i F k since σ i,k σ i,k. By this argument, it ollows that i is an element o all the sets F,..., F k. Furthermore, i i R k+hen the assignment to s k is non-zero, i.e. i F k R k. Again i has nonzero assignments or servers,... s k. Furthermore, the has always a non-zero assignment to the irst server, i.e. F R. Lemma states that Algorithm is correct or a single server. We prove the correctness by an inuction over the number o servers an assume that it computes a possible vali assignment or m servers. Assume that an assignment exists an that the algorithm oes not in one. Then, we istinguish two cases. Let n be the number o iles that can be assigne by the algorithm beore it aborts. ) For all i [n ], j [m ] : σ i,j > σ i,j+ : Consier the server s m. So, either σ n,m < or m j= n + < n +. In the later case no vali assignment is possible. The irst case is equivalent to n + i= a i,m > c m, where or all i R m F m (i.e. a i,m ) we have i F... F m an thereore m i a i,m = i j=

5 i. Further- Assume that a i,j Then, constraint () implies a i,j more j [m] a i,j = i an thus, is a vali assignment o all iles. m a i i,m i j= Constraint () implies n i= a i,m c m an thereore i R m F m a i,m c m which implies This contraicts i R m F m i m j= i R m F m a i,m > c m i c m. an thereore in this case Algorithm must in a vali assignment. ) There exists i [n ] an j [m ] : σ i,j = σ i,j+ We show that servers s j an s j+ can be replace by a single server s such that Algorithm chooses the very same assignment, i.e. i a i enotes the assignment o ile i to the new server s then a i = a i,j + a i,j+. Let l := min{i : j [m ] : σ i,j = σ i,j+ }. The capacity o the new server is c j + c j+ an the banwith is b = + c j+ an the sustainability σ = cj+cj+ ++ is thus σ j = c j+ + c j + c j+ + + c j = σ j+. So, the alternative set o servers are sorte as,,..., s j, s, s j+,..., s m accoring to their sustainability. So, server s simply replaces s j an s j+. We consier the ollowing cases an check whether Algorithm chooses the same assignment, i.e. a i = a i,j + a i,j+ without changing all other assignments by an inuction over i. a) i > l: Then, σ i,j = σ i,j+ an thereore i R j i R j+ an F j i F j+. Thereore rom Algorithm an the equivalency o both algorithms it ollows that a i = a i,j + a i,j+ since b = + +. b) i = l: Then i R j an R j+. Let j j an j j + be the participating servers at the event. Then or all l [j, j ]: a i,l = c i,l P = j :i F j b l j :i R j P, where t i + j :i R j c i,j i Since by inuction c i = c i,j + c i,j+ an b = + +, we have a i = a j + a j+. For the other assignments in the interval [j, j ] the values o not change either. c) i < l, i R j R j+ This case is impossible since otherwise σ i,j = σ i,j+ which leas to the contraiction i < l = min{i : σ i,j = σ i,j+ }. ) i < l, i (R j F j ) an (R j+ F j+ ) This case is impossible by the esign o the Algorithm. e) i < l, i R j an F j+ This case is also not possible. ) i < l, i (R j F j ) an (R j+ F j+ ) Then a i = a i,j = a i,j+ =. g) i < l, i (R j F j ) an (R j+ F j+ ) So, the rest o the ile is store on server s j (an alternatively server s ). Then a i,j+ = an a i,j = i t i j :i F j The very same ormula computes a j an thereore a j = a i,j + a i,j+ h) i < l an F j an R j+ I i R j+, then σ i,j+ = σ i,j+ which contraicts the minimality o l. Thereore, j + > m or i R j+ F j+. So, we get a i,j = i a i,j+ = i while a i = i j j = i j j = t i Thereore a i = a i,j + a i,j+. i) i < l, i F j an F j+ Then a i = i bj+bj+ = a i,j + a i,j+. By inuction over i it ollows that all assignments o all servers j < j an j > j + remain the same or both sets o servers. We have assume that an assignment (a i,j ) exists or the set o servers S = {,..., s n }. Then, there also exists an assignment or the set o servers S = {,..., s j, s, s j+,..., s n }, namely a i,k, k < j a i,k = a i,j + a i,j+, k = j a i,k+, k > j By the inuction hypothesis, Algorithm ins such an assignment. However, we have assume that Algorithm oes not compute an assignment or m+ storage evices. Then, we have prove that or the given iles the algorithm behaves ientical as in the case o m storage evices. So, it shoul also not compute an assignment or m storage evices which is a contraiction. So, it always ins a solution i it exists.

6 IV. CONCLUSION We have presente a solution or optimally istributing iles in a istribute storage system or meia streaming. Our Algorithm optimally istributes iles on in O(nm + m log m), thus is a vast improvement over an o ar programming solution. This allows storing iles with known banwith requirement to a istribute storage system such that reistributing ata ater the assignment is never necessary. REFERENCES [] Giuseppe DeCania, Deniz Hastorun, Maan Jampani, Gunavarhan Kakulapati, Avinash Lakshman, Alex Pilchin, Swaminathan Sivasubramanian, Peter Vosshall, an Werner Vogels. Dynamo: Amazon s highly available key-value store. SIGOPS Operating Systems Review, (6):5, 7. [] Peter Druschel an Antony I. T. Rowstron. Past: A large-scale, persistent peer-to-peer storage utility. In HotOS, pages 75 8,. [] Sanjay Ghemawat, Howar Gobio, an Shun-Tak Leung. The google ile system. In SOSP : Proceeings o the Nineteenth ACM Symposium on Operating Systems Principles, page9, New York, NY, USA,. ACM. [] Davi Karger, Eric Lehman, Tom Leighton, Rina Panigrahy, Matthew Levine, an Daniel Lewin. Consistent hashing an ranom trees: istribute caching protocols or relieving hot spots on the worl wie web. In STOC 97: Proceeings o the twenty-ninth annual ACM symposium on Theory o computing, pages 65 66, New York, NY, USA, 997. ACM. [5] John Kubiatowicz, Davi Binel, Yan Chen, Steven Czerwinski, Patrick Eaton, Dennis Geels, Ramakrishna Gummai, Sean Rhea, Hakim Weatherspoon, Chris Wells, an Ben Zhao. Oceanstore: An architecture or global-scale persistent storage. In ASPLOS-IX: Proceeings o the Ninth International Conerence on Architectural Support or Programming Languages an Operating Systems, page9, New York, NY, USA,. ACM. [6] Tobias Langner, Christian Schinelhauer, an Alexaner Souza. Optimal ile-istribution in heterogeneous an asymmetric storage networks. In Proceeings o the 7th international conerence on Current trens in theory an practice o computer science, SOFSEM, page68 8, Berlin, Heielberg,. Springer-Verlag. [7] Davi A. Patterson, Garth Gibson, an Rany H. Katz. A case or reunant arrays o inexpensive isks (rai). In SIGMOD 88: Proceeings o the 988 ACM SIGMOD International Conerence on Management o Data, page9 6, New York, NY, USA, 988. ACM. [8] Christian Schinelhauer an Gunnar Schomaker. Weighte istribute hash tables. In SPAA 5: Proceeings o the 7th ACM Symposium on Parallelism in Algorithms an Architectures, page8 7, Las Vegas, Nevaa, USA, 7 - July 5. ACM Press, New York, NY, USA. [9] Christian Schinelhauer an Gunnar Schomaker. SAN optimal multi parameter access scheme. In ICNICONSMCL 6: Proceeings o the International Conerence on Networking, International Conerence on Systems an International Conerence on Mobile Communications an Learning Technologies, page 8, Washington, DC, USA, 6. IEEE Computer Society. [] Steen Schott. Datenverteilung au heterogene speicher unter em atenstrom-ansatz. Master s thesis, Albert-Luwigs-Universität Freiburg, Germany,. [] Xiaohui Shen an Alok Chouhary. DPFS: A istribute parallel ile system. In ICPP : Proceeings o the International Conerence on Parallel Processing, page 5, Washington, DC, USA,. IEEE Computer Society. [] The Wuala Project. Wuala [On; accesse July ]. b b b b b b b b b b banwith s banwith s banwith s banwith s s stops ile ile ile ile ile Fig.. Sweep Line On Algorithm istributing ile onto servers,..., s.

Risk Management for Derivatives

Risk Management for Derivatives Risk Management or Derivatives he Greeks are coming the Greeks are coming! Managing risk is important to a large number o iniviuals an institutions he most unamental aspect o business is a process where

More information

Dynamo: Amazon s Highly Available Key-value Store

Dynamo: Amazon s Highly Available Key-value Store Dynamo: Amazon s Highly Available Key-value Store Giuseppe DeCandia, Deniz Hastorun, Madan Jampani, Gunavardhan Kakulapati, Avinash Lakshman, Alex Pilchin, Swaminathan Sivasubramanian, Peter Vosshall and

More information

A Data Placement Strategy in Scientific Cloud Workflows

A Data Placement Strategy in Scientific Cloud Workflows A Data Placement Strategy in Scientific Clou Workflows Dong Yuan, Yun Yang, Xiao Liu, Jinjun Chen Faculty of Information an Communication Technologies, Swinburne University of Technology Hawthorn, Melbourne,

More information

Game Theoretic Modeling of Cooperation among Service Providers in Mobile Cloud Computing Environments

Game Theoretic Modeling of Cooperation among Service Providers in Mobile Cloud Computing Environments 2012 IEEE Wireless Communications an Networking Conference: Services, Applications, an Business Game Theoretic Moeling of Cooperation among Service Proviers in Mobile Clou Computing Environments Dusit

More information

Enterprise Resource Planning

Enterprise Resource Planning Enterprise Resource Planning MPC 6 th Eition Chapter 1a McGraw-Hill/Irwin Copyright 2011 by The McGraw-Hill Companies, Inc. All rights reserve. Enterprise Resource Planning A comprehensive software approach

More information

A Generalization of Sauer s Lemma to Classes of Large-Margin Functions

A Generalization of Sauer s Lemma to Classes of Large-Margin Functions A Generalization of Sauer s Lemma to Classes of Large-Margin Functions Joel Ratsaby University College Lonon Gower Street, Lonon WC1E 6BT, Unite Kingom J.Ratsaby@cs.ucl.ac.uk, WWW home page: http://www.cs.ucl.ac.uk/staff/j.ratsaby/

More information

Firewall Design: Consistency, Completeness, and Compactness

Firewall Design: Consistency, Completeness, and Compactness C IS COS YS TE MS Firewall Design: Consistency, Completeness, an Compactness Mohame G. Goua an Xiang-Yang Alex Liu Department of Computer Sciences The University of Texas at Austin Austin, Texas 78712-1188,

More information

Algorithms and Methods for Distributed Storage Networks 9 Analysis of DHT Christian Schindelhauer

Algorithms and Methods for Distributed Storage Networks 9 Analysis of DHT Christian Schindelhauer Algorithms and Methods for 9 Analysis of DHT Institut für Informatik Wintersemester 2007/08 Distributed Hash-Table (DHT) Hash table does not work efficiently for inserting and deleting Distributed Hash-Table

More information

Modelling and Resolving Software Dependencies

Modelling and Resolving Software Dependencies June 15, 2005 Abstract Many Linux istributions an other moern operating systems feature the explicit eclaration of (often complex) epenency relationships between the pieces of software

More information

V:Drive - Costs and Benefits of an Out-of-Band Storage Virtualization System

V:Drive - Costs and Benefits of an Out-of-Band Storage Virtualization System V:Drive - Costs and Benefits of an Out-of-Band Storage Virtualization System André Brinkmann, Michael Heidebuer, Friedhelm Meyer auf der Heide, Ulrich Rückert, Kay Salzwedel, and Mario Vodisek Paderborn

More information

INFLUENCE OF GPS TECHNOLOGY ON COST CONTROL AND MAINTENANCE OF VEHICLES

INFLUENCE OF GPS TECHNOLOGY ON COST CONTROL AND MAINTENANCE OF VEHICLES 1 st Logistics International Conference Belgrae, Serbia 28-30 November 2013 INFLUENCE OF GPS TECHNOLOGY ON COST CONTROL AND MAINTENANCE OF VEHICLES Goran N. Raoičić * University of Niš, Faculty of Mechanical

More information

MSc. Econ: MATHEMATICAL STATISTICS, 1995 MAXIMUM-LIKELIHOOD ESTIMATION

MSc. Econ: MATHEMATICAL STATISTICS, 1995 MAXIMUM-LIKELIHOOD ESTIMATION MAXIMUM-LIKELIHOOD ESTIMATION The General Theory of M-L Estimation In orer to erive an M-L estimator, we are boun to make an assumption about the functional form of the istribution which generates the

More information

Scalable Multiple NameNodes Hadoop Cloud Storage System

Scalable Multiple NameNodes Hadoop Cloud Storage System Vol.8, No.1 (2015), pp.105-110 http://dx.doi.org/10.14257/ijdta.2015.8.1.12 Scalable Multiple NameNodes Hadoop Cloud Storage System Kun Bi 1 and Dezhi Han 1,2 1 College of Information Engineering, Shanghai

More information

10.2 Systems of Linear Equations: Matrices

10.2 Systems of Linear Equations: Matrices SECTION 0.2 Systems of Linear Equations: Matrices 7 0.2 Systems of Linear Equations: Matrices OBJECTIVES Write the Augmente Matrix of a System of Linear Equations 2 Write the System from the Augmente Matrix

More information

Ch 10. Arithmetic Average Options and Asian Opitons

Ch 10. Arithmetic Average Options and Asian Opitons Ch 10. Arithmetic Average Options an Asian Opitons I. Asian Option an the Analytic Pricing Formula II. Binomial Tree Moel to Price Average Options III. Combination of Arithmetic Average an Reset Options

More information

A REVIEW ON EFFICIENT DATA ANALYSIS FRAMEWORK FOR INCREASING THROUGHPUT IN BIG DATA. Technology, Coimbatore. Engineering and Technology, Coimbatore.

A REVIEW ON EFFICIENT DATA ANALYSIS FRAMEWORK FOR INCREASING THROUGHPUT IN BIG DATA. Technology, Coimbatore. Engineering and Technology, Coimbatore. A REVIEW ON EFFICIENT DATA ANALYSIS FRAMEWORK FOR INCREASING THROUGHPUT IN BIG DATA 1 V.N.Anushya and 2 Dr.G.Ravi Kumar 1 Pg scholar, Department of Computer Science and Engineering, Coimbatore Institute

More information

New Algorithms for Load Balancing in Peer-to-Peer Systems

New Algorithms for Load Balancing in Peer-to-Peer Systems New Algorithms for Load Balancing in Peer-to-Peer Systems David R. Karger Matthias Ruhl MIT Laboratory for Computer Science Cambridge, MA 02139, USA {karger, ruhl}@theory.lcs.mit.edu Abstract Load balancing

More information

Hull, Chapter 11 + Sections 17.1 and 17.2 Additional reference: John Cox and Mark Rubinstein, Options Markets, Chapter 5

Hull, Chapter 11 + Sections 17.1 and 17.2 Additional reference: John Cox and Mark Rubinstein, Options Markets, Chapter 5 Binomial Moel Hull, Chapter 11 + ections 17.1 an 17.2 Aitional reference: John Cox an Mark Rubinstein, Options Markets, Chapter 5 1. One-Perio Binomial Moel Creating synthetic options (replicating options)

More information

Cross-Over Analysis Using T-Tests

Cross-Over Analysis Using T-Tests Chapter 35 Cross-Over Analysis Using -ests Introuction his proceure analyzes ata from a two-treatment, two-perio (x) cross-over esign. he response is assume to be a continuous ranom variable that follows

More information

SOLVING LOAD REBALANCING FOR DISTRIBUTED FILE SYSTEM IN CLOUD

SOLVING LOAD REBALANCING FOR DISTRIBUTED FILE SYSTEM IN CLOUD International Journal of Advances in Applied Science and Engineering (IJAEAS) ISSN (P): 2348-1811; ISSN (E): 2348-182X Vol-1, Iss.-3, JUNE 2014, 54-58 IIST SOLVING LOAD REBALANCING FOR DISTRIBUTED FILE

More information

A Comparison of Performance Measures for Online Algorithms

A Comparison of Performance Measures for Online Algorithms A Comparison of Performance Measures for Online Algorithms Joan Boyar 1, Sany Irani 2, an Kim S. Larsen 1 1 Department of Mathematics an Computer Science, University of Southern Denmark, Campusvej 55,

More information

On Adaboost and Optimal Betting Strategies

On Adaboost and Optimal Betting Strategies On Aaboost an Optimal Betting Strategies Pasquale Malacaria 1 an Fabrizio Smerali 1 1 School of Electronic Engineering an Computer Science, Queen Mary University of Lonon, Lonon, UK Abstract We explore

More information

How To Connect Two Servers Together In A Data Center Network

How To Connect Two Servers Together In A Data Center Network DPillar: Scalable Dual-Port Server Interconnection for Data Center Networks Yong Liao ECE Department University of Massachusetts Amherst, MA 3, USA Dong Yin Automation Department Northwestern Polytech

More information

Seminar Presentation for ECE 658 Instructed by: Prof.Anura Jayasumana Distributed File Systems

Seminar Presentation for ECE 658 Instructed by: Prof.Anura Jayasumana Distributed File Systems Seminar Presentation for ECE 658 Instructed by: Prof.Anura Jayasumana Distributed File Systems Prabhakaran Murugesan Outline File Transfer Protocol (FTP) Network File System (NFS) Andrew File System (AFS)

More information

How To Price Internet Access In A Broaban Service Charge On A Per Unit Basis

How To Price Internet Access In A Broaban Service Charge On A Per Unit Basis iqui Pricing for Digital Infrastructure Services Subhajyoti Banyopahyay * an sing Kenneth Cheng Department of Decision an Information Sciences Warrington College of Business Aministration University of

More information

An Indexing Framework for Efficient Retrieval on the Cloud

An Indexing Framework for Efficient Retrieval on the Cloud An Indexing Framework for Efficient Retrieval on the Cloud Sai Wu National University of Singapore wusai@comp.nus.edu.sg Kun-Lung Wu IBM T. J. Watson Research Center klwu@us.ibm.com Abstract The emergence

More information

BOSCH. CAN Specification. Version 2.0. 1991, Robert Bosch GmbH, Postfach 30 02 40, D-70442 Stuttgart

BOSCH. CAN Specification. Version 2.0. 1991, Robert Bosch GmbH, Postfach 30 02 40, D-70442 Stuttgart CAN Specification Version 2.0 1991, Robert Bosch GmbH, Postfach 30 02 40, D-70442 Stuttgart CAN Specification 2.0 page 1 Recital The acceptance an introuction of serial communication to more an more applications

More information

How To Understand The Structure Of A Can (Can)

How To Understand The Structure Of A Can (Can) Thi t t ith F M k 4 0 4 BOSCH CAN Specification Version 2.0 1991, Robert Bosch GmbH, Postfach 50, D-7000 Stuttgart 1 The ocument as a whole may be copie an istribute without restrictions. However, the

More information

Distributed Data Stores

Distributed Data Stores Distributed Data Stores 1 Distributed Persistent State MapReduce addresses distributed processing of aggregation-based queries Persistent state across a large number of machines? Distributed DBMS High

More information

Secure Cache Provision: Provable DDOS Prevention for Randomly Partitioned Services with Replication

Secure Cache Provision: Provable DDOS Prevention for Randomly Partitioned Services with Replication 2013 IEEE 33rd International Conference on Distributed Computing Systems Workshops Secure Cache Provision: Provable DDOS Prevention for Randomly Partitioned Services with Replication Weibo Chu, Xiaohong

More information

Storage Systems Autumn 2009. Chapter 6: Distributed Hash Tables and their Applications André Brinkmann

Storage Systems Autumn 2009. Chapter 6: Distributed Hash Tables and their Applications André Brinkmann Storage Systems Autumn 2009 Chapter 6: Distributed Hash Tables and their Applications André Brinkmann Scaling RAID architectures Using traditional RAID architecture does not scale Adding news disk implies

More information

A Universal Sensor Control Architecture Considering Robot Dynamics

A Universal Sensor Control Architecture Considering Robot Dynamics International Conference on Multisensor Fusion an Integration for Intelligent Systems (MFI2001) Baen-Baen, Germany, August 2001 A Universal Sensor Control Architecture Consiering Robot Dynamics Frierich

More information

Sharding by Hash Partitioning A database scalability pattern to achieve evenly sharded database clusters

Sharding by Hash Partitioning A database scalability pattern to achieve evenly sharded database clusters Sharding by Hash Partitioning A database scalability pattern to achieve evenly sharded database clusters Caio H. Costa 1, João Vianney B. M. Filho 1, Paulo Henrique M. Maia 1, Francisco Carlos M. B. Oliveira

More information

A New Evaluation Measure for Information Retrieval Systems

A New Evaluation Measure for Information Retrieval Systems A New Evaluation Measure for Information Retrieval Systems Martin Mehlitz martin.mehlitz@ai-labor.e Christian Bauckhage Deutsche Telekom Laboratories christian.bauckhage@telekom.e Jérôme Kunegis jerome.kunegis@ai-labor.e

More information

JON HOLTAN. if P&C Insurance Ltd., Oslo, Norway ABSTRACT

JON HOLTAN. if P&C Insurance Ltd., Oslo, Norway ABSTRACT OPTIMAL INSURANCE COVERAGE UNDER BONUS-MALUS CONTRACTS BY JON HOLTAN if P&C Insurance Lt., Oslo, Norway ABSTRACT The paper analyses the questions: Shoul or shoul not an iniviual buy insurance? An if so,

More information

Data Center Power System Reliability Beyond the 9 s: A Practical Approach

Data Center Power System Reliability Beyond the 9 s: A Practical Approach Data Center Power System Reliability Beyon the 9 s: A Practical Approach Bill Brown, P.E., Square D Critical Power Competency Center. Abstract Reliability has always been the focus of mission-critical

More information

The one-year non-life insurance risk

The one-year non-life insurance risk The one-year non-life insurance risk Ohlsson, Esbjörn & Lauzeningks, Jan Abstract With few exceptions, the literature on non-life insurance reserve risk has been evote to the ultimo risk, the risk in the

More information

Minimum-Energy Broadcast in All-Wireless Networks: NP-Completeness and Distribution Issues

Minimum-Energy Broadcast in All-Wireless Networks: NP-Completeness and Distribution Issues Minimum-Energy Broacast in All-Wireless Networks: NP-Completeness an Distribution Issues Mario Čagal LCA-EPFL CH-05 Lausanne Switzerlan mario.cagal@epfl.ch Jean-Pierre Hubaux LCA-EPFL CH-05 Lausanne Switzerlan

More information

P2P Networking - Advantages and Disadvantages of Virtualization

P2P Networking - Advantages and Disadvantages of Virtualization Are Virtualized Overlay Networks Too Much of a Good Thing? Pete Keleher, Bobby Bhattacharjee, Bujor Silaghi Department of Computer Science University of Maryland, College Park keleher@cs.umd.edu 1 Introduction

More information

Optimal Control Policy of a Production and Inventory System for multi-product in Segmented Market

Optimal Control Policy of a Production and Inventory System for multi-product in Segmented Market RATIO MATHEMATICA 25 (2013), 29 46 ISSN:1592-7415 Optimal Control Policy of a Prouction an Inventory System for multi-prouct in Segmente Market Kuleep Chauhary, Yogener Singh, P. C. Jha Department of Operational

More information

Unbalanced Power Flow Analysis in a Micro Grid

Unbalanced Power Flow Analysis in a Micro Grid International Journal of Emerging Technology an Avance Engineering Unbalance Power Flow Analysis in a Micro Gri Thai Hau Vo 1, Mingyu Liao 2, Tianhui Liu 3, Anushree 4, Jayashri Ravishankar 5, Toan Phung

More information

View Synthesis by Image Mapping and Interpolation

View Synthesis by Image Mapping and Interpolation View Synthesis by Image Mapping an Interpolation Farris J. Halim Jesse S. Jin, School of Computer Science & Engineering, University of New South Wales Syney, NSW 05, Australia Basser epartment of Computer

More information

Products no longer available

Products no longer available Technical ata sheet R6..R haracterize control valves, 2-way, with flange PN 6 for open an close col an warm water systems for moulating control on the water sie of air-hanling an heating systems air bubble-tight

More information

Lecture L25-3D Rigid Body Kinematics

Lecture L25-3D Rigid Body Kinematics J. Peraire, S. Winall 16.07 Dynamics Fall 2008 Version 2.0 Lecture L25-3D Rigi Boy Kinematics In this lecture, we consier the motion of a 3D rigi boy. We shall see that in the general three-imensional

More information

Load Balancing for Heterogeneous Web Servers

Load Balancing for Heterogeneous Web Servers Loa Balancing for Heterogeneous Web Servers Aam Piórkowski 1, Aleksaner Kempny 2, Arian Hajuk 1, an Jacek Strzelczyk 1 1 Department of Geoinfomatics an Applie Computer Science, AGH University of Science

More information

MODELLING OF TWO STRATEGIES IN INVENTORY CONTROL SYSTEM WITH RANDOM LEAD TIME AND DEMAND

MODELLING OF TWO STRATEGIES IN INVENTORY CONTROL SYSTEM WITH RANDOM LEAD TIME AND DEMAND art I. robobabilystic Moels Computer Moelling an New echnologies 27 Vol. No. 2-3 ransport an elecommunication Institute omonosova iga V-9 atvia MOEING OF WO AEGIE IN INVENOY CONO YEM WIH ANOM EA IME AN

More information

Cost Efficient Datacenter Selection for Cloud Services

Cost Efficient Datacenter Selection for Cloud Services Cost Efficient Datacenter Selection for Clou Services Hong u, Baochun Li henryxu, bli@eecg.toronto.eu Department of Electrical an Computer Engineering University of Toronto Abstract Many clou services

More information

Achieving quality audio testing for mobile phones

Achieving quality audio testing for mobile phones Test & Measurement Achieving quality auio testing for mobile phones The auio capabilities of a cellular hanset provie the funamental interface between the user an the raio transceiver. Just as RF testing

More information

Design and Implementation of Performance Guaranteed Symmetric Load Balancing Algorithm

Design and Implementation of Performance Guaranteed Symmetric Load Balancing Algorithm Design and Implementation of Performance Guaranteed Symmetric Load Balancing Algorithm Shaik Nagoor Meeravali #1, R. Daniel *2, CH. Srinivasa Reddy #3 # M.Tech, Department of Information Technology, Vignan's

More information

ISSN: 2277-3754 ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 12, June 2014

ISSN: 2277-3754 ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 12, June 2014 ISSN: 77-754 ISO 900:008 Certifie International Journal of Engineering an Innovative echnology (IJEI) Volume, Issue, June 04 Manufacturing process with isruption uner Quaratic Deman for Deteriorating Inventory

More information

Stochastic Planning for Content Delivery: Unveiling the Benefits of Network Functions Virtualization

Stochastic Planning for Content Delivery: Unveiling the Benefits of Network Functions Virtualization Stochastic Planning for Content Delivery: Unveiling the Benefits of Network Functions Virtualization Michele Mangili, Fabio Martignon an Antonio Capone LRI, Université Paris-Su {michele.mangili, fabio.martignon}@lri.fr

More information

Web Appendices of Selling to Overcon dent Consumers

Web Appendices of Selling to Overcon dent Consumers Web Appenices of Selling to Overcon ent Consumers Michael D. Grubb A Option Pricing Intuition This appenix provies aitional intuition base on option pricing for the result in Proposition 2. Consier the

More information

Improving Direct Marketing Profitability with Neural Networks

Improving Direct Marketing Profitability with Neural Networks Volume 9 o.5, September 011 Improving Direct Marketing Profitability with eural etworks Zaiyong Tang Salem State University Salem, MA 01970 ABSTRACT Data mining in irect marketing aims at ientifying the

More information

Bellini: Ferrying Application Traffic Flows through Geo-distributed Datacenters in the Cloud

Bellini: Ferrying Application Traffic Flows through Geo-distributed Datacenters in the Cloud Bellini: Ferrying Application Traffic Flows through Geo-istribute Datacenters in the Clou Zimu Liu, Yuan Feng, an Baochun Li Department of Electrical an Computer Engineering, University of Toronto Department

More information

Big Data Racing and parallel Database Technology

Big Data Racing and parallel Database Technology EFFICIENT DATA ANALYSIS SCHEME FOR INCREASING PERFORMANCE IN BIG DATA Mr. V. Vivekanandan Computer Science and Engineering, SriGuru Institute of Technology, Coimbatore, Tamilnadu, India. Abstract Big data

More information

Calibration of the broad band UV Radiometer

Calibration of the broad band UV Radiometer Calibration of the broa ban UV Raiometer Marian Morys an Daniel Berger Solar Light Co., Philaelphia, PA 19126 ABSTRACT Mounting concern about the ozone layer epletion an the potential ultraviolet exposure

More information

MANAGEMENT OF DATA REPLICATION FOR PC CLUSTER BASED CLOUD STORAGE SYSTEM

MANAGEMENT OF DATA REPLICATION FOR PC CLUSTER BASED CLOUD STORAGE SYSTEM MANAGEMENT OF DATA REPLICATION FOR PC CLUSTER BASED CLOUD STORAGE SYSTEM Julia Myint 1 and Thinn Thu Naing 2 1 University of Computer Studies, Yangon, Myanmar juliamyint@gmail.com 2 University of Computer

More information

A Study on Workload Imbalance Issues in Data Intensive Distributed Computing

A Study on Workload Imbalance Issues in Data Intensive Distributed Computing A Study on Workload Imbalance Issues in Data Intensive Distributed Computing Sven Groot 1, Kazuo Goda 1, and Masaru Kitsuregawa 1 University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan Abstract.

More information

Consumer Referrals. Maria Arbatskaya and Hideo Konishi. October 28, 2014

Consumer Referrals. Maria Arbatskaya and Hideo Konishi. October 28, 2014 Consumer Referrals Maria Arbatskaya an Hieo Konishi October 28, 2014 Abstract In many inustries, rms rewar their customers for making referrals. We analyze the optimal policy mix of price, avertising intensity,

More information

Dynamic Network Security Deployment Under Partial Information

Dynamic Network Security Deployment Under Partial Information Dynamic Network Security Deployment Uner Partial nformation nvite Paper) George Theoorakopoulos EPFL Lausanne, Switzerlan Email: george.theoorakopoulos @ epfl.ch John S. Baras University of Marylan College

More information

Detecting Possibly Fraudulent or Error-Prone Survey Data Using Benford s Law

Detecting Possibly Fraudulent or Error-Prone Survey Data Using Benford s Law Detecting Possibly Frauulent or Error-Prone Survey Data Using Benfor s Law Davi Swanson, Moon Jung Cho, John Eltinge U.S. Bureau of Labor Statistics 2 Massachusetts Ave., NE, Room 3650, Washington, DC

More information

ThroughputScheduler: Learning to Schedule on Heterogeneous Hadoop Clusters

ThroughputScheduler: Learning to Schedule on Heterogeneous Hadoop Clusters ThroughputScheuler: Learning to Scheule on Heterogeneous Haoop Clusters Shehar Gupta, Christian Fritz, Bob Price, Roger Hoover, an Johan e Kleer Palo Alto Research Center, Palo Alto, CA, USA {sgupta, cfritz,

More information

zupdate: Updating Data Center Networks with Zero Loss

zupdate: Updating Data Center Networks with Zero Loss zupate: Upating Data Center Networks with Zero Loss Hongqiang Harry Liu Yale University hongqiang.liu@yale.eu Lihua Yuan Microsoft lyuan@microsoft.com Xin Wu Duke University xinwu@cs.uke.eu Roger Wattenhofer

More information

A New Pricing Model for Competitive Telecommunications Services Using Congestion Discounts

A New Pricing Model for Competitive Telecommunications Services Using Congestion Discounts A New Pricing Moel for Competitive Telecommunications Services Using Congestion Discounts N. Keon an G. Ananalingam Department of Systems Engineering University of Pennsylvania Philaelphia, PA 19104-6315

More information

Optimizing Multiple Stock Trading Rules using Genetic Algorithms

Optimizing Multiple Stock Trading Rules using Genetic Algorithms Optimizing Multiple Stock Traing Rules using Genetic Algorithms Ariano Simões, Rui Neves, Nuno Horta Instituto as Telecomunicações, Instituto Superior Técnico Av. Rovisco Pais, 040-00 Lisboa, Portugal.

More information

Inverse Trig Functions

Inverse Trig Functions Inverse Trig Functions c A Math Support Center Capsule February, 009 Introuction Just as trig functions arise in many applications, so o the inverse trig functions. What may be most surprising is that

More information

Optimal Energy Commitments with Storage and Intermittent Supply

Optimal Energy Commitments with Storage and Intermittent Supply Submitte to Operations Research manuscript OPRE-2009-09-406 Optimal Energy Commitments with Storage an Intermittent Supply Jae Ho Kim Department of Electrical Engineering, Princeton University, Princeton,

More information

Scalable live video streaming to cooperative clients using time shifting and video patching

Scalable live video streaming to cooperative clients using time shifting and video patching calable live vieo streaming to cooperative clients using time shifting an vieo patching Meng Guo, Mostafa H. Ammar {mguo, ammar}@cc.gatech.eu Networking an Telecommunication Group ollege of omputing, Georgia

More information

Digital barrier option contract with exponential random time

Digital barrier option contract with exponential random time IMA Journal of Applie Mathematics Avance Access publishe June 9, IMA Journal of Applie Mathematics ) Page of 9 oi:.93/imamat/hxs3 Digital barrier option contract with exponential ranom time Doobae Jun

More information

A Highly Scalable and Efficient Distributed File Storage System

A Highly Scalable and Efficient Distributed File Storage System IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.12, December 2009 101 A Highly Scalable and Efficient Distributed File Storage System Fawad Riasat Raja Dr. Adeel Akram Faculty

More information

Improving Emulation Throughput for Multi-Project SoC Designs

Improving Emulation Throughput for Multi-Project SoC Designs Improving Emulation Throhput for Multi-Project SoC Designs By Frank Schirrmeister, Caence Design Systems As esign sizes grow, so, too, oes the verification effort. Inee, verification has become the biggest

More information

Minimizing Makespan in Flow Shop Scheduling Using a Network Approach

Minimizing Makespan in Flow Shop Scheduling Using a Network Approach Minimizing Makespan in Flow Shop Scheuling Using a Network Approach Amin Sahraeian Department of Inustrial Engineering, Payame Noor University, Asaluyeh, Iran 1 Introuction Prouction systems can be ivie

More information

Low-Complexity and Distributed Energy Minimization in Multi-hop Wireless Networks

Low-Complexity and Distributed Energy Minimization in Multi-hop Wireless Networks Low-Complexity an Distribute Energy inimization in ulti-hop Wireless Networks Longbi Lin, Xiaojun Lin, an Ness B. Shroff Center for Wireless Systems an Applications (CWSA) School of Electrical an Computer

More information

Answers to the Practice Problems for Test 2

Answers to the Practice Problems for Test 2 Answers to the Practice Problems for Test 2 Davi Murphy. Fin f (x) if it is known that x [f(2x)] = x2. By the chain rule, x [f(2x)] = f (2x) 2, so 2f (2x) = x 2. Hence f (2x) = x 2 /2, but the lefthan

More information

Parameterized Algorithms for d-hitting Set: the Weighted Case Henning Fernau. Univ. Trier, FB 4 Abteilung Informatik 54286 Trier, Germany

Parameterized Algorithms for d-hitting Set: the Weighted Case Henning Fernau. Univ. Trier, FB 4 Abteilung Informatik 54286 Trier, Germany Parameterize Algorithms for -Hitting Set: the Weighte Case Henning Fernau Trierer Forschungsberichte; Trier: Technical Reports Informatik / Mathematik No. 08-6, July 2008 Univ. Trier, FB 4 Abteilung Informatik

More information

Joining Cassandra. Luiz Fernando M. Schlindwein Computer Science Department University of Crete Heraklion, Greece mattos@csd.uoc.

Joining Cassandra. Luiz Fernando M. Schlindwein Computer Science Department University of Crete Heraklion, Greece mattos@csd.uoc. Luiz Fernando M. Schlindwein Computer Science Department University of Crete Heraklion, Greece mattos@csd.uoc.gr Joining Cassandra Binjiang Tao Computer Science Department University of Crete Heraklion,

More information

The higher education factor: The role of higher education in the hiring and promotion practices in the fire service. By Nick Geis.

The higher education factor: The role of higher education in the hiring and promotion practices in the fire service. By Nick Geis. The higher eucation factor: The role of higher eucation in the hiring an promotion practices in the fire service. By Nick Geis Spring 2012 A paper submitte to the faculty of The University of North Carolina

More information

Mathematical Models of Therapeutical Actions Related to Tumour and Immune System Competition

Mathematical Models of Therapeutical Actions Related to Tumour and Immune System Competition Mathematical Moels of Therapeutical Actions Relate to Tumour an Immune System Competition Elena De Angelis (1 an Pierre-Emmanuel Jabin (2 (1 Dipartimento i Matematica, Politecnico i Torino Corso Duca egli

More information

An Introduction to Event-triggered and Self-triggered Control

An Introduction to Event-triggered and Self-triggered Control An Introuction to Event-triggere an Self-triggere Control W.P.M.H. Heemels K.H. Johansson P. Tabuaa Abstract Recent evelopments in computer an communication technologies have le to a new type of large-scale

More information

A MPCP-Based Centralized Rate Control Method for Mobile Stations in FiWi Access Networks

A MPCP-Based Centralized Rate Control Method for Mobile Stations in FiWi Access Networks A MPCP-Based Centralized Rate Control Method or Mobile Stations in FiWi Access Networks 215 IEEE. Personal use o this material is permitted. Permission rom IEEE must be obtained or all other uses, in any

More information

State of Louisiana Office of Information Technology. Change Management Plan

State of Louisiana Office of Information Technology. Change Management Plan State of Louisiana Office of Information Technology Change Management Plan Table of Contents Change Management Overview Change Management Plan Key Consierations Organizational Transition Stages Change

More information

GPRS performance estimation in GSM circuit switched services and GPRS shared resource systems *

GPRS performance estimation in GSM circuit switched services and GPRS shared resource systems * GPRS performance estimation in GSM circuit switche serices an GPRS share resource systems * Shaoji i an Sen-Gusta Häggman Helsinki Uniersity of Technology, Institute of Raio ommunications, ommunications

More information

The most common model to support workforce management of telephone call centers is

The most common model to support workforce management of telephone call centers is Designing a Call Center with Impatient Customers O. Garnett A. Manelbaum M. Reiman Davison Faculty of Inustrial Engineering an Management, Technion, Haifa 32000, Israel Davison Faculty of Inustrial Engineering

More information

Distributed Metadata Management Scheme in HDFS

Distributed Metadata Management Scheme in HDFS International Journal of Scientific and Research Publications, Volume 3, Issue 5, May 2013 1 Distributed Metadata Management Scheme in HDFS Mrudula Varade *, Vimla Jethani ** * Department of Computer Engineering,

More information

A performance analysis of EtherCAT and PROFINET IRT

A performance analysis of EtherCAT and PROFINET IRT A perormance analysis o EtherCAT and PROFINET IRT Conerence paper by Gunnar Prytz ABB AS Corporate Research Center Bergerveien 12 NO-1396 Billingstad, Norway Copyright 2008 IEEE. Reprinted rom the proceedings

More information

Data Management Course Syllabus

Data Management Course Syllabus Data Management Course Syllabus Data Management: This course is designed to give students a broad understanding of modern storage systems, data management techniques, and how these systems are used to

More information

Sensor Network Localization from Local Connectivity : Performance Analysis for the MDS-MAP Algorithm

Sensor Network Localization from Local Connectivity : Performance Analysis for the MDS-MAP Algorithm Sensor Network Localization from Local Connectivity : Performance Analysis for the MDS-MAP Algorithm Sewoong Oh an Anrea Montanari Electrical Engineering an Statistics Department Stanfor University, Stanfor,

More information

ANALYSIS OF SMART METER DATA USING HADOOP

ANALYSIS OF SMART METER DATA USING HADOOP ANALYSIS OF SMART METER DATA USING HADOOP 1 Balaji K. Bodkhe, 2 Dr. Sanjay P. Sood MESCOE Pune, CDAC Mohali Email: 1 balajibodkheptu@gmail.com, 2 spsood@gmail.com Abstract The government agencies and the

More information

Distributed File Systems

Distributed File Systems Distributed File Systems Paul Krzyzanowski Rutgers University October 28, 2012 1 Introduction The classic network file systems we examined, NFS, CIFS, AFS, Coda, were designed as client-server applications.

More information

R.Tamilarasi #1, G.Kesavaraj *2

R.Tamilarasi #1, G.Kesavaraj *2 ENHANCING SECURE MULTI USER ACCESS IN CLOUD ENVIRONMENT BY LOAD BALANCING RTamilarasi #1, GKesavaraj *2 #1 Mphil, Research Scholar, Vivekananda Arts and Science College for women *2 Assistant professor,department

More information

Stock Market Value Prediction Using Neural Networks

Stock Market Value Prediction Using Neural Networks Stock Market Value Preiction Using Neural Networks Mahi Pakaman Naeini IT & Computer Engineering Department Islamic Aza University Paran Branch e-mail: m.pakaman@ece.ut.ac.ir Hamireza Taremian Engineering

More information

LONG TERM EVOLUTION WITH 5G USING MAPREDUCING TASK FOR DISTRIBUTED FILE SYSTEMS IN CLOUD

LONG TERM EVOLUTION WITH 5G USING MAPREDUCING TASK FOR DISTRIBUTED FILE SYSTEMS IN CLOUD LONG TERM EVOLUTION WITH 5G USING MAPREDUCING TASK FOR DISTRIBUTED FILE SYSTEMS IN CLOUD 1 MSSoundarya, 2 GSiva Kumar Assistant Professor, Department of CSE Gnanamani College of Engineering ABSTRACT -

More information

Search Advertising Based Promotion Strategies for Online Retailers

Search Advertising Based Promotion Strategies for Online Retailers Search Avertising Base Promotion Strategies for Online Retailers Amit Mehra The Inian School of Business yeraba, Inia Amit Mehra@isb.eu ABSTRACT Web site aresses of small on line retailers are often unknown

More information

Option Pricing for Inventory Management and Control

Option Pricing for Inventory Management and Control Option Pricing for Inventory Management an Control Bryant Angelos, McKay Heasley, an Jeffrey Humpherys Abstract We explore the use of option contracts as a means of managing an controlling inventories

More information

DECISION SUPPORT SYSTEM FOR MANAGING EDUCATIONAL CAPACITY UTILIZATION IN UNIVERSITIES

DECISION SUPPORT SYSTEM FOR MANAGING EDUCATIONAL CAPACITY UTILIZATION IN UNIVERSITIES DECISION SUPPORT SYSTEM OR MANAGING EDUCATIONAL CAPACITY UTILIZATION IN UNIVERSITIES Svetlana Vinnik 1, Marc H. Scholl 2 Abstract Decision-making in the fiel of acaemic planning involves extensive analysis

More information

11 CHAPTER 11: FOOTINGS

11 CHAPTER 11: FOOTINGS CHAPTER ELEVEN FOOTINGS 1 11 CHAPTER 11: FOOTINGS 11.1 Introuction Footings are structural elements that transmit column or wall loas to the unerlying soil below the structure. Footings are esigne to transmit

More information

Math 230.01, Fall 2012: HW 1 Solutions

Math 230.01, Fall 2012: HW 1 Solutions Math 3., Fall : HW Solutions Problem (p.9 #). Suppose a wor is picke at ranom from this sentence. Fin: a) the chance the wor has at least letters; SOLUTION: All wors are equally likely to be chosen. The

More information

Multi-objective Optimization with Fuzzy Based Ranking for TCSC Supplementary Controller to Improve Rotor Angle and Voltage Stability

Multi-objective Optimization with Fuzzy Based Ranking for TCSC Supplementary Controller to Improve Rotor Angle and Voltage Stability International Journal o lectrical an lectronics ngineering 3:9 9 Multi-obective Optimization with Fuzzy Base Ranking or CSC Supplementary Controller to Improve Rotor Angle an oltage Stability S. Pana,

More information

Web Appendices to Selling to Overcon dent Consumers

Web Appendices to Selling to Overcon dent Consumers Web Appenices to Selling to Overcon ent Consumers Michael D. Grubb MIT Sloan School of Management Cambrige, MA 02142 mgrubbmit.eu www.mit.eu/~mgrubb May 2, 2008 B Option Pricing Intuition This appenix

More information

CALCULATION INSTRUCTIONS

CALCULATION INSTRUCTIONS Energy Saving Guarantee Contract ppenix 8 CLCULTION INSTRUCTIONS Calculation Instructions for the Determination of the Energy Costs aseline, the nnual mounts of Savings an the Remuneration 1 asics ll prices

More information