Franck Diard
Chief Software Architect at NVIDIA

Franck Diard

GeForce NOW, a Cloud Gaming Service by NVIDIA

Abstract: NVIDIA has built a global cloud gaming service to stream PC game titles from the cloud to its SHIELD mobile and console products, launched in May 2015. This presentation will explain some of the technical challenges and how they were met. Topics will include hosting GPU servers in the cloud with virtualization, how to deliver the snappiest graphics for the best user experience, video compression, resource optimization, and will be concluded by a live demo.

Biography: Franck Diard is Chief Software Architect at NVIDIA. In 16 years at NVIDIA, he has started, designed, built and helped deliver significant PC technologies like SLI, Optimus hybrid graphics and GRID cloud graphics platforms. In 2007, he was named the first software Distinguished Engineer at NVIDIA in recognition of his outstanding contributions and engineering creativity, with 70 US patents granted so far. He holds a Ph.D. in Computer Science from University of Nice-Sophia Antipolis, France.

Jan Van Looy
Assistant Professor at the research group for Media and ICT (iMinds-MICT), Ghent University, Belgium
with support from Anissa All
& Elena Nunez Castellar

Jan Van Looy

Serious NetGames? Evaluating the potential of network technologies for serious games and learning assessment

Abstract: With the maturing of tools for creating networked games, their development costs are decreasing, bringing them within reach of lower-cost productions aimed at digital game-based learning (DGBL). This opens new horizons for enhancing gameplay, implementing learning strategies and realizing new concepts altogether. Areas for potential improvement include adaptive game content, social reward strategies and personal monitoring. Moreover, network technologies create new possibilities for assessing the effectiveness of DGBL itself. Until recently, a clear conceptualization and standardized procedure to asses DGBL effectiveness were lacking. For the past three years, we have been developing such a framework and conducting validation studies in a variety of domains: training mental arithmetic skills and teaching English to young children; teaching traffic safety to teenagers, fire safety to hospital staff and interpersonal communication to the personnel of a large European bank. In my keynote I will present some of the lessons learned whilst conducting this work and reflect upon both achievements and failures. Based on these I will assess the potential of network technologies for furthering the impact of serious games and their effectiveness assessment.

Biography: Jan Van Looy is Assistant Professor at the research group for Media and ICT (iMinds-MICT), Ghent University, Belgium. In 2006 he finished his Ph.D. on the shaping of digital games, which was published as Understanding Computer Game Culture (2010). In 2007-2008, he worked as a Postdoctoral Fellow at HUMLab, University of Umea, Sweden. Since September 2008, he has been working, first as Postdoctoral Researcher and then as Assistant Professor at Ghent University. There he and his team conduct multidisciplinary user research into immersive media and digital gaming. Past publications have dealt with the effects of stereoscopic 3D media, omnidirectional video experience, measuring player identification; antecedents of gamer identity; effects of stereotype threat on game experience, teacher adoption of digital games in the classroom, and effectiveness of serious games for language learning, awareness raising, mental calculus and personal empowerment and social inclusion. Current research interests of the Gaming and Immersive Media Lab ( include brain correlates of flow experience, psychophysiological effects of gender stereotype threat, methodology for serious game effectiveness research, player persistence in exergames, and augmented reality quality of experience.

Maxim Baryshnikov
Solutions Architect at

Maxim Baryshnikov

Engineering decisions behind World of Tanks server

Abstract: World of Tanks is MMOG with recorded peak of simultaneously online players of more then 1.1 million. "Authoritarian" server architecture (all world processing including physics calculation, movement, shooting, damage absorption, etc. happens on server side and being then dictated to clients) provides great defense against cheating, but also greatly rises the pressure on servers. This session looks into engineering decisions which made it possible for World of Tanks to scale out, particularly following challenges and solutions will be covered:
  • Decoupling server components and distributing them among servers in one datacenter keeping in mind redundancy and fail-over recovery questions. As a result a bullet which has been fired on one physical server may hit the target on another one.
  • Employing effective udp-based core network protocol, which is being used both for communication between server components and server-to-clients data exchange. This allows almost linear server scaling in one datacenter, so only network throughput of the datacenter becomes scaling limit.
  • Scaling out of one datacenter using cluster-of-clusters approach. Goals: scalability beyond one datacenter, better player experience through lower latency while preserving world integrity for a player.
  • Another challenge was how to deal with large amount of players data on server (as of early 2015 only players dossiers exceeds 500Gb of compressed data of about 30 million active players of Russia & CIS region). Data exchange and processing solution was based on messaging queues, transferring and processing more then 3000 events per second. This also involves building data streams for Business Intelligence and real-time reaction on certain event chains.
  • Large amount of distributed data raises the issue of data consistency among different components and services. We solve this by designing all our components with principles of "self-healing" of inconsistent data, using both reliable and unreliable data message queues depending on data nature, "visible consistency" of even inconsistent data, real-time monitoring.
All these challenges brought World of Tanks to highly distributed component based architecture of the server itself which in it's turn is tightly integrated with large-scale service oriented "platform" infrastructure: web portals, commerce services, public APIs, ratings, clans, cyber sports logic services, etc., more than 70 services in total. I also plan to briefly cover our plans for further improvements of our service platform solution like strictly defining it's parts and responsibilities, data flows, more flexibility in rule-based business decisions made real-time, etc.

Biography: Maksim Baryshnikov is a solutions Architect at He has graduated from Belaruski Nacyjanal'ny Tehnieny Universitet in 2007. Key expertise areas he has built in his 12 years of software engineering and architecture experience are software systemˇ¦s performance, scalability and reliability. Heˇ¦s been working on improving informational services infrastructure of legendary World of Tanks game as well as making robust and flexible SOA foundation for future titles. Currently technically leads development of new generation of Clan Wars ˇX turn-based land domination strategy game involving World of Tanks gameplay.ˇ¨

Contacts: Maha Abdallah and Mirko Suznjevic (Workshop Co-Chairs)

Organized by:


Copyright © 2015 Maha Abdallah & Mirko Suznjevic
NetGames 2015 Logo designed by Frederic Delvalle