Intelligent Placement of Datacenter for Internet Services
Upcoming SlideShare
Loading in...5
×
 

Intelligent Placement of Datacenter for Internet Services

on

  • 279 views

Presentation for EEDC Master course, UPC

Presentation for EEDC Master course, UPC

Statistics

Views

Total Views
279
Views on SlideShare
272
Embed Views
7

Actions

Likes
0
Downloads
5
Comments
0

1 Embed 7

http://web.ict.kth.se 7

Accessibility

Upload Details

Uploaded via as Adobe PDF

Usage Rights

CC Attribution-NonCommercial-ShareAlike LicenseCC Attribution-NonCommercial-ShareAlike LicenseCC Attribution-NonCommercial-ShareAlike License

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

Intelligent Placement of Datacenter for Internet Services Intelligent Placement of Datacenter for Internet Services Presentation Transcript

  • Overview Framework Description Exploring Tradeoffs Future Work . . Intelligent placement of DCs for Internet Services . .. . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo I˜ Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} FIB UPC May 7, 2012 . . . . . .by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work . . . Overview 1 . . . Framework Description 2 Parameters The Optimization Problem & Solutions Exploring the tool Evaluating Solutions . . . Exploring Tradeoffs 3 . . . Future Work 4 . . . . . .by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Datacenters Explosion . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Location . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. The problem . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. A solution Provide a framework to automate location selection to satisfy: Customers Needs Low Company Costs Friendliness to environment . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions . . . Overview 1 . . . Framework Description 2 Parameters The Optimization Problem & Solutions Exploring the tool Evaluating Solutions . . . Exploring Tradeoffs 3 . . . Future Work 4 . . . . . .by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. Costs COSTS = CAPEX + OPEX OPEX i.e.: CAPEX i.e.: maintenance & land administration electricity coverage personnel network coverage electricity & water usage . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. More Response Time Consistency Delay Availability CO2 Emisions All parameters depend directly to location and may contradict each other. . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. Problem Use parameters to formulate a Non-Linear cost model and optimization problem. Inputs: Outputs: Total max number of servers Optimal Cost Expected average utilization Max numbers of servers Users per Server Number of servers that serve Redundancy Level each population center Max network delay & consistency delay Min availability CAPEX and OPEX cost ∀ locations . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. Solutions HARD Problem (non-linear,large number of location). Need for custom solutions. Simple linear programming (LP0) Pre-set linear programming (LP1) Brute Force (Brute) Heuristic based on LP (Heuristic) Simulated Annealing plus LP1 (SA+LP1) Optimized SA+LP1 (OSA+LP1) . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. Input Data Location-dependent: Datacenter Characteristics: . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. Characterizing Locations Figure: Datacenter cost depending on location Figure: Location characteristics . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. Was it worth it? Figure: Cost per month distribution in 500 locations of the US Figure: Cost of a network of datacenters. for 60K servers . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Parameters Framework Description The Optimization Problem & Solutions Exploring Tradeoffs Exploring the tool Future Work Evaluating Solutions. Best Algorithm? OSA+LP1 is the best solution. . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work . . . Overview 1 . . . Framework Description 2 Parameters The Optimization Problem & Solutions Exploring the tool Evaluating Solutions . . . Exploring Tradeoffs 3 . . . Future Work 4 . . . . . .by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Solutions Cost vs Latency Golden ratio: 50 ms . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Solutions Cost vs Latency Golden ratio: 50 ms Cost vs Availability Cheaper if less redundant Golden ratio: Tier II . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Solutions Cost vs Latency Golden ratio: 50 ms Cost vs Availability Cheaper if less redundant Golden ratio: Tier II Cost vs Consistency speed Consistency and latency are conflicting . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Solutions Cost vs Latency Golden ratio: 50 ms Cost vs Availability Cheaper if less redundant Golden ratio: Tier II Cost vs Consistency speed Consistency and latency are conflicting Cost vs Environment for max latencies >70ms: green network $100K more expensive per month . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Solutions Cost vs Latency Golden ratio: 50 ms Cost vs Availability Cheaper if less redundant Golden ratio: Tier II Cost vs Consistency speed Consistency and latency are conflicting Cost vs Environment for max latencies >70ms: green network $100K more expensive per month Cost vs Chillers for max latencies >70ms: Avoiding chillers ⇒ ↓ cost 8% . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work . . . Overview 1 . . . Framework Description 2 Parameters The Optimization Problem & Solutions Exploring the tool Evaluating Solutions . . . Exploring Tradeoffs 3 . . . Future Work 4 . . . . . .by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services
  • Overview Framework Description Exploring Tradeoffs Future Work. Improvements How can we improve the framework? Automatic data collection Frequency of data updates Covering other parts of the world . . . . . . by `nigo Goiri, Kien Le, Jordi Guitart, Jordi Torres, and Ricardo Bianchini Presentor: Emmanouil Dimogerontakis @{EEDC} I˜ Intelligent placement of DCs for Internet Services