4. Why cloud?
Common development issues
Cases
Storage and processing
Multiplayer
Mobile apps
Analytics and big data
DevOps
Some experience
5. Why cloud?
Games become services
Approach to delivery the games is changing
Support and maintenance is changing
Monetization is changing constantly and becoming “live” –
business model of one project constantly adapting to the market
situation
Studios and indie developers getting the access to the new
markets
8. Case 1: storage
Performance counters data can be collected every second
Mobile app: every user action = new event
Multiplayer: amount of msgs/sec can be >100 000
We need a big, scalable, fast and flexible storage
19. Azure Mobile Services
Windows Store
iOS
Android
Xamarin
Windows Phone 8
iOS
Android
HTML5/JS
Open Source SDK
Source
Control
Scripting:
table,
scheduled
& custom
API
REST API
Facebook Twitter Microsoft Google
Active
Directory
SQL
Table
Storage
Blob
Storage
WNS & MPNS APNS GCM
Mongo
DB
22. Blobs
Binary content
SQL Azure
SQL Server as a
service
Virtual Machines
Windows/Linux virtual
machines for heavy
workloads
Web Apps
Admin panel/users
portal
23. Case 4: predictive analytics and big data
Halo 4 - Hadoop HDInsight for getting
insights from raw data
Insights = which modes are preferred
by gamers; session of the length;
amount of kills, etc
24. Case 4: predictive analytics and big data
HDInsight
Hadoop
clusters on
demand
ML
As a service
Stream
Analytics
NRT data
streaming
and/or
processing
Data
Factory
Data
transformati
on
Batch
HPC as a
service
25. Ситуация 5: DevOps
iPhone app does not
work
Slow requests
DB is not
responding Latency
Campaign is not
popular
27. Some experience
Cloud-based games is a good approach
But for doing things right sometimes you need to:
optimize the requests lifecycle
optimize the format of the messages
choose right instruments
testing proactively
doing the system loosely coupled