: Massivizing Online Games using Cloud Computing Alexandru Iosup Parallel and Distributed Systems Group, Delft University of Technology, The Netherlands VENI Massivizing Online Games using Cloud Computing
Cloud Computing VENI Massivizing Online Games using Cloud Computing
Joe’s Small Online Gaming Studio Variability is Joe’s unsolved problem Predictor +IT Source ($) days Own (10M$/iter) Cloud (1$-1M$) Static Own (100M$) Year # of Users Exponential growth VENI Massivizing Online Games using Cloud Computing My research:
The Scientific Challenge Devise a software and hardware architecture that can tolerate high variability without service degradation, just by adding more resources. Grow and shrink infrastructure automatically Replace faulty parts automatically Automation is efficient One support engineer regardless of scale I’ve tackled similar problems for grids and P2P systems VENI Massivizing Online Games using Cloud Computing Variability is one of the few unsolved problems in Computer Science
Variability = Player + Non-Player Interaction ? ? ? VENI Massivizing Online Games using Cloud Computing Research Challenge
New IT Fabric = Cloud + New Variability-Aware Methods ? VENI Massivizing Online Games using Cloud Computing Research Challenge
What if Joe is not into Gaming? e-Science e-Collaboration e-Architecture VENI Massivizing Online Games using Cloud Computing
8 Extra Slides
Why Am I the Person to Work on This?
Why Should the Netherlands Fund This? Gaming is entertainment branch with highest revenue Source: The Economist, Benchmarking IT industry competitiveness Human Capital: NL is 27 th + “A longer-term challenge for some European countries is encouraging more graduates to choose science- related subjects.” but
VENI Massivizing Online Games using Cloud Computing 11 Sources: MMOGChart, own research.Sources: ESA, MPAA, RIAA. Gaming, the New World Market: 7BN$/year 250,000,000+ players 10,000+ online games Success kills small companies Only big companies can join market (1BN+) Exponential growth Stop innovating and the growth stops tomorrow!
VENI Massivizing Online Games using Cloud Computing 12 MSG, MMOG, MMO, … 1.Virtual world + 2.Game data Player activity, stats, and (social) relationships Romeo and Juliet Massivized Online Games = online games with massive numbers of players (100K+), for which the social interaction improves the experience “Data deluge” = New variability source up to 40% revenueup to 40% operation costs Old variability source
Is Lack of Data a Risk? (Reviewer #2) Situation world-wide “Unfortunately, I am not able to release any of the data. Sony still considers it proprietary, and we can't control that.” – Sony “releases” EverQuest data “Unfortunately, CCP won't allow me to make their data publicly available for EVE Online.” – CCP “releases” EVE Online data My situation: NO Risk Spent three years collecting data for gaming and clouds KGS, OpenTTD, RuneScape, WoW, AmazonWS, GoogleAE, RightScale Prior to this, spent five years collecting data for grids, P2P, general The Grid Workloads Archive 6 traces online, over 20 in total Toolset for analysis/statistical modeling
Why Near-Real-Time Gaming Analytics? (Reviewer #2) Situation world-wide More detailed/faster analysis is constant concern “Data deluge” in E-Science etc. TNO thinks it’s big Online Gaming Three years ago gaming analytics was not a market Last year Zynga made ~700M$, mainly from it Playtomic, Kontangent, Mochibot: using own infrastructure, reduced refresh time from weeks to hours (minutes, early 2011) Zillians thinks it’s big