Back to overview

Cosmmus: an infrastructure for massively multiplayer online games

English title Cosmmus: an infrastructure for massively multiplayer online games
Applicant Pedone Fernando
Number 146404
Funding scheme Project funding
Research institution Istituto di sistemi informatici (SYS) Facoltà di scienze informatiche
Institution of higher education Università della Svizzera italiana - USI
Main discipline Information Technology
Start/End 01.04.2013 - 30.09.2016
Approved amount 176'325.00
Show all

Keywords (3)

distributed systems; scalability; high availability

Lay Summary (Italian)

Cosmmus: una infrastruttura per Massively Multiplayer Online Games (MMOGs)
Lay summary
I videogiochi online multigiocatore per un pubblico di massa (Massively multiplayer online games - MMOGs) sono diventati un genere importante di intrattenimento degli ultimi dieci anni. In questi giochi, molte migliaia di partecipanti giocano contemporaneamente uno con l'altro nello stesso ambiente virtuale. Questo tipo di applicazione è stata tradizionalmente sostenuta da infrastrutture centrali potenti e costose, dotate di elevate capacità di calcolo e di accesso alla Rete. Questo progetto si propone di esaminare alcune delle componenti fondamentali di una infrastruttura geograficamente distribuita a supporto del concetto MMOG. Le caratteristiche più importanti di tale infrastruttura sono la coerenza di sistema, le prestazioni e la resistenza ai guasti.

Nonostante la nostra applicazione principale sia MMOG, ci aspettiamo che i protocolli sviluppati in questo progetto siano utilizzabili in settori più ampi, tra cui ambienti virtuali in rete, come i motori di simulazione distribuita. L'applicabilità dei nostri risultati non è limitata agli ambienti virtuali. Per esempio, meccanismi di controllo di coerenza a latenza efficiente e protocolli di comunicazione di gruppo, potrebbero essere utilizzati per migliorare le prestazioni delle banche dati distribuite. Infine, qualsiasi applicazione che usufruisse di funzionalità distribuite tra più computer e che richiedesse un elevato controllo di coerenza, potrebbero trarre beneficio dal nostro lavoro.
Direct link to Lay Summary Last update: 28.03.2013

Responsible applicant and co-applicants



Dynamic Scalable State Machine Replication
Hoang Long, Bezerra Eduardo, Pedone Fernando (2016), Dynamic Scalable State Machine Replication, in IEEE/IFIP International Conference on Dependable Systems and Networks (DSN), IEEE, IEEE.
Strong Consistency at Scale
Bezerra Eduardo, Hoang Long, Pedone Fernando (2016), Strong Consistency at Scale, in Bulletin of the IEEE Computer Society Technical Committee on Data Engineering, 39(1), 93.
Ridge: high-throughput, low-latency atomic multicast
Bezerra Eduardo, Cason Daniel, Pedone Fernando (2015), Ridge: high-throughput, low-latency atomic multicast, in Reliable Distributed Systems (SRDS), 2015 IEEE 34th Symposium on, IEEE, IEEE.
Rethinking State-Machine Replication for Parallelism
Marandi Parisa, Bezerra Eduardo, Pedone Fernando (2014), Rethinking State-Machine Replication for Parallelism, in IEEE 34th International Conference on Distributed Computing Systems, IEEE, IEEE.
Scalable State-Machine Replication
Bezerra Eduardo, Pedone Fernando, van Renesse Robbert (2014), Scalable State-Machine Replication, in Dependable Systems and Networks (DSN), 2014 44th Annual IEEE/IFIP International Conference on.


Group / person Country
Types of collaboration
Prof. B. Garbinato, UNIL Switzerland (Europe)
- Publication
Prof. W. Zwaenepoel, EPFL Switzerland (Europe)
- Exchange of personnel

Associated projects

Number Title Start Funding scheme
169066 Cosmmus 2: an infrastructure for scalable distributed applications 01.10.2016 Project funding


Massively multiplayer online games (MMOGs) have become an important genre of online entertainment in the last decade. In these games, many thousands of participants play simultaneously with one another in the same virtual environment. This kind of application has been traditionally supported by powerful and expensive central infrastructures with considerable processor and network capabilities. This project intends to look into some of the key components of a geographically distributed infrastructure to support MMOGs. The most important features in such an infrastructure are consistency, performance, and fault tolerance.MMOGs are characterized by a large number of players. To support so many participants, a client-server infrastructure is generally used, where client requests are handled by a cluster of servers, and sometimes a datacenter. In this paradigm, the servers have the official copy of the game state and are responsible for computing state changes over time; clients are responsible for presenting the game state to the players. Servers are also responsible for updating the clients in real time on the changes that occur in the game virtual environment. Because of the high cost to maintain a centralized infrastructure, alternative solutions have considered distributing communication and processing among participants of a game, following a peer-to-peer model. Solutions of this kind, however, come with their own design challenges. Two of the main challenges are performance and fairness. Performance is mostly related to ensuring low response time for player actions, guaranteeing a good game experience. Fairness is concerned with malicious players who could try to subvert the game engine for their benefit, for example violating the rules of the game (i.e., cheating). While centralized infrastructures also have to deal with such problems, decentralized environments are more vulnerable to them.The typical solution to boost performance (i.e., reduce latency) in MMOGs is to relax consistency. Inconsistencies in a game can be tolerated but are usually annoying (e.g., an object that suddenly disappears from a location and magically reappears in another). As we elaborate in the proposal, our ongoing work has considered optimistic techniques to reduce the latency of distributed game engines, with the advantage of preserving consistency. Work on decentralized MMOGs is too incipient to have addressed the problem of fairness. We discuss some approaches we intend to pursue to prevent and detect cheating.While our motivating application is MMOGs, we expect the protocols designed in this project to be applicable in broader areas. The closest one is networked virtual environments which are not games, such as applications for military training or flight simulation. But the applicability of our results is not limited to virtual environments. Latency-efficient consistency mechanisms and group communication protocols may be used to improve performance of distributed databases, for example. Finally, any application that holds a state divided among several computers and requires strong consistency may benefit from our work.