Free-scaling your Data Center

Free-scaling your Data Center

The increasing popularity of both small and large private clouds and expanding public clouds poses new requirements to data center (DC) architectures. First, DC architectures should be incrementally scalable, i.e., the architecture should support the creation of DCs of arbitrary size while retaining key performance characteristics. Second, initial DC deployments should be incrementally expandable, i.e., the architecture should support small-scale upgrades without a notable decrease in operation efficiency. A DC architecture possessing both properties satisfies the requirement of free-scaling. Recent work in DC design focuses on traditional performance and scalability characteristics, therefore resulting in symmetric topologies whose upgradability is coarse-grained at best. In our earlier work we proposed Scafida, an asymmetric, scale-free network inspired DC topology which scales incrementally and has favorable structural characteristics. In this paper, we build on Scafida and propose a full-fledged DC architecture achieving free-scaling. Our main contribution is threefold. First, we propose an organic expansion algorithm for Scafida; this combined with Scafida’s flexible original design results in a freely scalable architecture. Second, we introduce the Effective Source Routing mechanism that provides near-shortest paths, multi-path and multicast capability and low signaling overhead by exploiting the benefits of the Scafida topology. Third, we show based on extensive simulations and a prototype implementation that Scafida is capable of handling the traffic patterns characteristic of both enterprise and cloud data centers, tolerates network equipment failures to a high degree, and allows for high bisection bandwidth.

Gulyás András, PhD, egyetemi adjunktus,  BME, Távközlési és Médiainformatikai Tanszék

Kutatási beszámoló 

Készítették: Sonkoly Balázs, egyetemi adjunktus BME, Távközlési és Médiainformatikai Tanszék, Gulyás András, Gyarmati László, Tuan A.Trinh, Biczók Gergely

A teljes kutatásibeszámoló letölthető innen (PDF)

2011. augusztus