Research On Dynamic Load Balancing Mechanism Based On Energy Aware In MMOG Network

In recent years,due to the emergence of smart phones and the rapid development of 4G mobile networks,the number of users of online games(ie,playing games via the Internet)has grown tremendously.MMOG(Massive Multiplayer Online Game)is an online game.In the virtual world of MMOG,each user has a representation in the game.This expression is called Avatar in the game and Avatar is under user control.Typically,there are thousands of people online in the virtual world,and some popular games can even reach 10,000 people online at the same time.Therefore,in order to resolve the large load changes in the game server cluster,many MMOG operators have deployed their gaming services on the cloud in order to take advantage of cloud computing,such as reduced operating costs,dynamic scalability,and on-demand use.The demand of a game server for resources(CPU,memory,network bandwidth)is related to how many players in the game world and what these players are doing.Therefore,insufficient allocation of game resources will lead to excessive delays,which will result in a decrease in the player's gaming experience and loss of players.In order to ensure that the players have a good gaming experience,a plenty of game operators now use excess resources to guarantee the player's gaming experience.However,over-allocation will result in a decrease in resource utilization,leading to increased energy consumption and a dramatic increase in the operating costs of game operators.In order to solve the high energy consumption problem of over-allocation,this paper proposes an energy-aware dynamic load balancing mechanism for MMOG networks.First,predict the number of players in the future game world through the neural network,calculate the load according to the number of players and give the game how much resources are needed,and then calculate how many virtual machines(VM)are needed according to the required resources while guaranteeing the player's experience,and then use the genetic algorithm to optimize the energy consumption as the goal to evenly distribute these VMs to physical servers on the cloud.Experiments show that the energy-aware dynamic load balancing method proposed in this paper is relative to the over-allocation method,which ensures the player's experience while reducing considerable energy consumption.