Energy-efficient Multi-tier Web Server Clusters

1
Energy-efficient Multi-tier Web Server Clusters

• Daniel Mosse

2
Introduction

• Practical use of power control of multi-tier
  clusters
• Application e-commerce or other datacenter
  usages
• Goal meet deadlines AND minimize energy
• Open software and commodity hardward

3
Cluster Model
4
Power Model

• CPUs can vary frequency and voltage
• The strategy is if there is time, slow down
  processor (and procesing)?
• Real-time tasks have deadlines (that is, tasks
  must finish by a certain time)?
• How much, then, can we slow down tasks?

5
Static speed adjustment
Assumption all tasks have the same deadline.
Smax
Smin
time
Select the speed based on worst-case execution
time,WCET, and deadline
6
Workload Generation

• TPC-W Benchmark for e-commerce (Transaction
  Processing Performance Council)?
• Multiple on-line browser sessions
• Dynamic page generation and database access
• Authentication through SSL
• Real-time properties

7
Measuring end-to-end time delay
Server node 1
Server node 2
Server node 3
Client
Front-end
req.php
req.php?r123
Img1.jpg?123
Img2.gif?123
Img3.jpg?123
img1.jpg?123
img2.gif?123
img3.jpg?123
8
Using a Controller

• Main metric Quality of Service
• Web interaction
• Web interaction response time (WIRT)?
• Controller metricTardiness of a web interaction
• Tardiness show more precise information about the
  completion of requests and resource usage

9
Control servers on-off also!

• If there are N machines, and we can put some
  machines to sleep, and turn some machines off
• How many machines are put to sleep?
• How many machines are turned off?
• When do we change state of machines?
• For machines that are on, how to set speed?
• Multi-core systems
• Heterogeneous vs homogeneous cores
• Change voltage for each core? What about shared
  memory?