Adaptive TTL schemes for load balancing of distributed web servers

With ever increasing web traffic, a distributed Web system can provide scalability and flexibility to cope with growing client demands. Load balancing algorithms to spread the load across multiple Web servers are crucial to achieve the scalability. Various domain name server (DNS) based schedulers have been proposed in the literature, mainly for multiple homogeneous servers. DNS provides (logical) host name to IP-address mapping (i.e., the server assignment), but the mapping is not done for each server access. This is because the address mapping is cached for a time-to-live (TTL) period to reduce network traffic. The presence of heterogeneous Web servers not only increases the complexity of the DNS scheduling problem, but also makes previously proposed algorithms for homogeneous distributed systems such as round robin not directly applicable. This leads us to propose new policies, called adaptive TTL algorithms, that take both the uneven distribution of client request rates and heterogeneity of Web servers into account to adaptively set the TTL value for each address mapping request. Extensive simulation results show that these strategies are effective in balancing load among geographically distributed heterogeneous Web servers.