EEDC                          34330                                   Intelligent Placement ofExecution                   ...
OUTLINE1. Introduction2. Example Datacenter3. Problem4. Placement of Datacenters5. Propose   5.1. Defining Framework   5.2...
Introduction    Internet services reach the whole world.    Millions of clients on the world.    Demand high availabili...
Example - Datacenter    Facebook - Prineville, Oregon USA        – 147,000-square-foot facility        – $200 million - $2...
Problem    Clients                ... widespreaded geographically                ... demand high availablity          ...
Problem    Clients                ... widespreaded geographically                ... demand high availablity          ...
Placement of DatacenterDirect impact on ...          Response time                  High availablity                  M...
OUTLINE1. Introduction2. Example Datacenter3. Problem4. Placement of Datacenters5. Propose   5.1. Defining Framework   5.2...
ProposeDatacenter automation of palcement of data centers.. Selection and selection and automation,efficiently !!        ...
Propose – Defining Framework    Parameters              Costs         •           CAPEX (Capital)           bringing ele...
Propose – Formulation    Subject to            Minimizing CAPEX and OPEX    Constraints            Response times < MA...
Propose – Solving    Problem is              ... non linear.              ... not directly solvable by Linear Programmi...
Tool       http://www.darklab.rutgers.edu/DCL/dcl.html                           13
Conclusion    No other work for intelligent placement of datacenters.    Contributions:              A framework is pro...
Upcoming SlideShare
Loading in...5
×

Umit hw6

133

Published on

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
133
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
2
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Umit hw6

  1. 1. EEDC 34330 Intelligent Placement ofExecution Datacenters for InternetEnvironments for ServicesDistributedComputingMaster in Computer Architecture,Networks and Systems - CANS Homework number: 6 Umit Cavus Buyuksahin umit.cavus.buyuksahin@ac.upc.edu
  2. 2. OUTLINE1. Introduction2. Example Datacenter3. Problem4. Placement of Datacenters5. Propose 5.1. Defining Framework 5.2. Formulation 5.3. Solving the problem6. Conclusion 2
  3. 3. Introduction Internet services reach the whole world. Millions of clients on the world. Demand high availabilityin short response time. Thus huge datacenters constructedaround the world They have many servers,cooling systems, energy power systems.. 3
  4. 4. Example - Datacenter Facebook - Prineville, Oregon USA – 147,000-square-foot facility – $200 million - $215 million.* http://www.oregonlive.com/business/index.ssf/2010/01/facebook_picks_prineville 4
  5. 5. Problem Clients  ... widespreaded geographically  ... demand high availablity  ... in short response time Many servers requirement. Supplying Energy Cooling system Building and operating datacenters Green Energy 5
  6. 6. Problem Clients  ... widespreaded geographically  ... demand high availablity  ... in short response time Many servers requirement. Supplying Energy Cooling system Building and operating datacenters Green Energy PLACEMENT OF DATACENTER !! 6
  7. 7. Placement of DatacenterDirect impact on ...  Response time  High availablity  Mirrored Datacenters  Closest one serves  Capital and Operational Costs  Land acquisition and building  Bring network and electricy  Electricity & Water  Staff  CO2 emmisions (indirect) 7
  8. 8. OUTLINE1. Introduction2. Example Datacenter3. Problem4. Placement of Datacenters5. Propose 5.1. Defining Framework 5.2. Formulation 5.3. Solving the problem6. Conclusion 8
  9. 9. ProposeDatacenter automation of palcement of data centers.. Selection and selection and automation,efficiently !! 9
  10. 10. Propose – Defining Framework Parameters  Costs • CAPEX (Capital) bringing electricity and network land and construction power, backup, cooling equipment • OPEX (Operational) maintaince and administor electrcicity and water price  Response Time • Latency & number of servers  Consistency Delay • Latency from mirrored datacenters  Availablity • #9 changes in each tier  CO2 emissions 10
  11. 11. Propose – Formulation Subject to  Minimizing CAPEX and OPEX Constraints  Response times < MAX LATENCY , ∀ users  Min consistency delay between 2 DCs < MAX DELAY  Min system availability > MIN AVAILABILITY Output  # of servers at each location  Minimized cost 11
  12. 12. Propose – Solving Problem is  ... non linear.  ... not directly solvable by Linear Programming. Linear Programming (LP) for potential solution. Simulated Annealing (SA) for consiring neighborings. CA + LP for cost optimization. Quality of results compared with Brute solution. Tool is built  ... automatic dacenter location selection  ... new parameters and constraints can be added 12
  13. 13. Tool http://www.darklab.rutgers.edu/DCL/dcl.html 13
  14. 14. Conclusion No other work for intelligent placement of datacenters. Contributions:  A framework is proposed by defining parameters  Based on parameters, optimization problem defined  Proposed the most efficient and accurate solution approach  A tool is built to automate location selection Experimental results shows  Millions dollar are saved 14
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×