Theoretical Computer Science ( IF 1.1 ) Pub Date : 2020-07-08 , DOI: 10.1016/j.tcs.2020.06.033 Tal Navon , David Peleg
We study the mixed approach to fault tolerance in the general context of server assignment in networks. The approach is based on mixing two different existing strategies, namely, reinforcement and backup. The former strategy protects clients by reinforcing the servers assigned to them and making them fault-resistant, possibly at a high cost, while the latter protects clients by assigning to them alternate low price backup servers that can replace their primary servers in case those fail. Applying the mixed approach to fault tolerance gives rise to new fault tolerant variations of known server assignment problems. We introduce several NP-hard problems of this type, including the mixed fault-tolerant dominating set problem, the mixed fault-tolerant centers problem, and the mixed fault-tolerant facility location problem, and present polynomial time approximation algorithms for them, demonstrating the viability of the mixed strategy for server assignment problems.
中文翻译:
服务器分配中的混合容错能力:增强和备份相结合
我们在网络中服务器分配的一般情况下研究混合容错方法。该方法基于混合两种不同的现有策略,即加固和备份。前一种策略通过增强分配给它们的服务器并使其具有抗故障性(可能以较高的成本)来保护客户端,而后一种策略则通过向其分配备用的低价备用服务器来保护客户端,这些备用价格低的备用服务器可以在发生故障时替换其主服务器。将混合方法应用于容错会导致已知服务器分配问题的新容错变体。我们介绍了这种类型的NP难问题,包括混合容错控制集问题,混合容错中心问题和混合容错设施位置问题,并针对它们提出了多项式时间近似算法,证明了混合策略解决服务器分配问题的可行性。