OnCall

1
OnCall

• Defeating Traffic Spikes with a Free-Market
  Application Cluster

James Norris Keith Coleman Armando Fox
George Candea Stanford University
2
Motivation
3
CNN.com

• September 11, 2001
• 4x traffic in a single day8x traffic on second
  day
• Offline for 2.5 hours, diminished service
  afterwards
• Forced to borrow servers from sister AOL-TW
  websites

337.4 M
162.4 M
Page Views
40 M
4
Slashdot, etc

• Slashdot Effect
• Knocks out sites (often at the worst possible
  time)
• Variable Traffic
• Ticket Sales
• Contests
• Online Fashion Shows
• etc

5
What to do?
6
One Option Overprovision

• Works for steady state fluctuations (but is
  it optimal?)
• Too expensive for spike conditions (8x
  servers for CNN)
• Think about it Like having a fixed size buffer
• Can only support 1000 entries ? Lame
• Stanford Axess Sorry, 49 people already
  logged in
• And in steady state there is so much waste
• So what do we do? Use dynamic allocation

7
What is OnCall?

• OnCall is
• a cluster management system designed to
  multiplex several (possibly competing) dynamic
  web applications onto a single cluster.
• Goal
• Make spike handling possible while providing
  useful resource guarantees to all apps

8
OnCall Overview

• Marketplace of Applications
• Applications rent and lend computing resources
  according to pre-defined market policies
• Generic Platform
• Based on VMs
• ? application generic
• ? fast app swapping

9
Marketplace
10
Market Rounds

• Offline
• Each application assigned ownership of G
  computers at a fixed price (or rate)
• Online
• Determine market equilibrium price, P, by
  querying each application
• Calculate new allocation sizes at price P
• Adjust allocations, moving computers from sellers
  to buyers
• Repeat every time quantum, t

11
Offline Market G

• G
• Each app owns G nodes
• Resource guarantees
• Never have to sell no matter what the price or
  what other apps demands, an app is guaranteed
  use of its G nodes
• Can lend by choice (if there are renters at
  desired price)
• Can rent extra nodes (if it needs to and/or can
  afford to)

12
Online Market
7 5 2 14, but I only have 10 nodes!
5 3 2 10 Perfect!
10 nodes in cluster
Marketplace
Policy
Policy
Policy
13
Online Market Policies

• Inputs

Output of computers desired at price P
POLICY
Price P
14
Example Market Policy
n lt G (no spike)

• For each round, application A computes the number
  of nodes, n, it needs to handle current traffic
• Ex Application A has a price threshold of 6
• If (P lt 6), A will ask for n nodes
• If (P 6), A will only ask for min(n, G) nodes
  it cant afford to rent extras

n gt G (spike)
15
Finding the Equilibrium

• Sample points along the different policy
  functions
• Determine the price at which the total number of
  nodes desired by all apps equals the total number
  of nodes available on the cluster

16
Notes and Assumptions

• Homogeneity Assumption
• Cluster is assumed to be homogeneousall nodes
  rented at same price (for simplicity)
• Swapping Costs
• Time delay cost in start up / shut down of an
  app on a node.
• If a rental contract is renewed, app runs on
  same node.
• P Only for Extras
• Apps only pay price P for nodes above and beyond
  their own G
• Ex Using 40, G 30
• ? 40 30 10 nodes at price P

17
Platform
18
Platform Overview
19
Runtime Operation

• Runtime cycle repeats every t
• Marketplace calculates equilibrium price (and
  thus application allocations)
• Managers assigns apps to physical nodes
  (minimizing shutdowns and startups)
• Manager signals Responders to shutdown and start
  new app, as necessary
• At end of round, Manager gathers new usage stats
  reports stats to Market Policies
• Repeat

20
Does this work?
21
Simulation Testbed

• Three Simulations, Four Traits
• Spike handling under unconstrained resources
• Spike handling under constrained resources
• Resource guarantees
• Fast server activation
• U.C. Berkeley X Cluster
• 30 Nodes (double CNN.com)
• Dual 1 GHz PIII, 1.5 GB RAM
• VMware GSX Server on Linux

22
Sim 1 Spike Handling

• G 10 for both apps
• App 1 handles spikes, App 2 makes
• Notice Lag time between node assigned ? node
  active

23
Sim 2 Resource Constraints

• G1 12, G2 6, G3 12
• App 1 has higher budget than App 2, but both
  spike
• App 1 handles spikes, App 2 sees guarantee, App 3
  makes
• App 2 buys more when App 1s spike subsides

24
Sim 3 Fast Activation
Platform OnCall Optimal OnCall Limited Standard with OS Standard w/out OS
Time until Active (s) 5-10 50-120 270-330 710-750

• OnCall Optimal Load VMs from suspended state
• OnCall Limited Load VMs from shutdown state
• Standard with OS OS already installed on node
• Standard without OS Must install OS first
• Significance
• Worst case, gt 2x improvement
• When spike lasts only 30 minutes, this is
  significant
• If you can startup quickly, accurate predictor is
  not critical

25
More on Markets
26
Marketplace Optimality

• What is optimal?
• Under resource constraints, those applications
  with the most utility to derive from the use of
  additional nodes are given those nodes
• Utility Curves
• Curve specifies dollar value an application
  derives from possessing a certain number of nodes
  for a specific time quantum.

Trivially Utility curves are always
monotonically non-decreasing (i.e. it is never
worse to own more nodes at a given total cost)
To be optimal Marginal utility curves are
always monotonically non-increasing (i.e. every
additional node is worth same or less than one
before)
27
Marketplace Fairness

• Markets are optimal if
• they are free and fair
• Anti-competitive behavior
• Monopoly/Oligopoly
• Aggressive tactics
• Fairness through Regulation
• Ensure enough distinct owners ? no monopoly
• Fine or ban app that engages in overtly
  anti-competitive behavior

28
Competitive vs Cooperative

• Competitive Environments
• Ex ASP, where app owners may be in competition
• Cooperative Environments
• Ex Search engine, Yahoogle
• Quick Case Study
• App 1 Paid web search (very high value in low
  latency)
• App 2 Ad-supported web search (high value in
  low latency)
• App 3 Crawler (latency OK, starvation not)
• For each app, model utility of running at a
  given time
• Benefit If you add an app, just need to model
  that app, not remodel whole system

29
Profit Through Efficiency

• Shut Down App
• ASP shuts down servers when it can buy them for
  less than the cost of keeping them running (A/C,
  utilities, etc)
• ASP can then add additional capacity and sell
  only when profitable

30
Future Work
31
Future Work

• VM caching
• Cache VMs to local disk (speculatively or as
  read from NAS)
• Fault tolerance
• Add master-backup fault tolerance to the OnCall
  Manager
• Performance statistics
• Provide market policies with additional
  statistics (e.g. end-to-end response time)
• Scalable data layer
• Add support for scalable persistent stores that
  would allow replication on the data tier.
• Multiplexing
• Study trade-offs of running several applications
  on one node

32
Questions?