Building the Halo 4
Services with Orleans
Qcon London 2015
Slide 2
Slide 2 text
Caitie McCaffrey!
Distributed Systems Engineer
@Caitie
CaitieM.com
Slide 3
Slide 3 text
No content
Slide 4
Slide 4 text
No content
Slide 5
Slide 5 text
Presence
Statistics
Title Files
Cheat Detection
User Generated Content
Slide 6
Slide 6 text
No content
Slide 7
Slide 7 text
No content
Slide 8
Slide 8 text
No content
Slide 9
Slide 9 text
No content
Slide 10
Slide 10 text
Halo:CE - 6.43 million
Halo 2 - 8.49 million
Halo 3 - 11.87 million
Halo 3: ODST - 6.22 million
Halo Reach - 9.52 million
Slide 11
Slide 11 text
$220 million in sales
!
1 million players online
Day One
Slide 12
Slide 12 text
$300 million in sales
!
4 million players online
!
31.4 million hours
Week One
Slide 13
Slide 13 text
11.6 million players
!
1.5 billion games
!
270 million hours
Overall
Slide 14
Slide 14 text
Architectural
Challenges
Slide 15
Slide 15 text
Load Patterns
Load Patterns
Slide 16
Slide 16 text
Azure Worker Roles
Azure Table
Azure Blob
Azure Service Bus
Slide 17
Slide 17 text
Always Available
Slide 18
Slide 18 text
Low Latency &
High Concurrency
Slide 19
Slide 19 text
Stateless 3 Tier !
Architecture
Slide 20
Slide 20 text
Latency Issues
Slide 21
Slide 21 text
Add A Cache
Slide 22
Slide 22 text
Concurrency
Issues
Slide 23
Slide 23 text
Data Locality
Slide 24
Slide 24 text
The Actor Model
A framework & basis for reasoning about concurrency
A Universal Modular Actor Formalism for Artificial Intelligence !
Carl Hewitt, Peter Bishop, Richard Steiger (1973)
Slide 25
Slide 25 text
Send A Message
Create a New Actor
Change Internal State
Slide 26
Slide 26 text
State-full Services
Slide 27
Slide 27 text
Philip A. Bernstein, Sergey Bykov, Alan Geller, Gabriel
Kliot, Jorgen Thelin
Orleans: Distributed Virtual
Actors for Programmability
and Scalability
eXtreme Computing Group MSR
Slide 28
Slide 28 text
“Orleans is a runtime and
programming model for
building distributed systems,
based on the actor model”
Slide 29
Slide 29 text
Virtual Actors
“An Orleans actor always exists, virtually. It
cannot be explicitly created or destroyed”
Slide 30
Slide 30 text
Virtual Actors
• Perpetual Existence
• Automatic Instantiation
• Location Transparency
• Automatic Scale out
Slide 31
Slide 31 text
Runtime
• Messaging
• Hosting
• Execution
Slide 32
Slide 32 text
Messaging
Guarantees
Slide 33
Slide 33 text
Let’s Talk About CAP
Slide 34
Slide 34 text
Orleans is AP
Slide 35
Slide 35 text
Programming Model
• .NET Framework!
• Actor Interfaces!
• Promises!
• Actor References
• Turns
• Persistence
Slide 36
Slide 36 text
No content
Slide 37
Slide 37 text
Programming Model
• .NET Framework
• Actor Interfaces
• Promises
• Actor References!
• Turns
• Persistence
Slide 38
Slide 38 text
No content
Slide 39
Slide 39 text
Programming Model
• .NET Framework
• Actor Interfaces
• Promises
• Actor References
• Turns!
• Persistence
Slide 40
Slide 40 text
No content
Slide 41
Slide 41 text
Reliability
“Orleans manages all aspects of reliability automatically”
Slide 42
Slide 42 text
TOO!
Slide 43
Slide 43 text
No content
Slide 44
Slide 44 text
TOO!
Slide 45
Slide 45 text
TOO!
Slide 46
Slide 46 text
Orleans & Halo
Slide 47
Slide 47 text
No content
Slide 48
Slide 48 text
Halo 4: Statistics
Service
Slide 49
Slide 49 text
Player Grain
Slide 50
Slide 50 text
Game Grain
Slide 51
Slide 51 text
No content
Slide 52
Slide 52 text
Performance &
Scalability
Slide 53
Slide 53 text
“Orleans applications run at very high CPU
Utilization. We have run load tests with full
saturation of 25 servers for many days at 90%+
CPU utilization without any instability”