PDLB: Path Diversity-aware Load Balancing with Adaptive Granularity in Data Center Network

Modern data center topologies often take the form of a multi-rooted tree with rich parallel paths to provide high bandwidth. However, various path diversities caused by traffic dynamics, link failures and heterogeneous switching equipments widely exist in production data center network. Therefore, the multi-path load balancer in data center should be robust to these diversities. Although prior fine-grained schemes such as RPS and Presto make full use of available paths, they are prone to experi-ence packet reordering problem under asymmetric topology. The coarse-grained solutions such as ECMP and LetFlow effectively avoid packet reordering, but easily lead to under-utilization of multiple paths. To cope with these inefficiencies, we propose a load balancing mechanism called PDLB, which adaptively adjusts flowcell granularity according to path diversity. PDLB increases flowcell granularity to alleviate packet reordering under large degrees of topology asymmetry, while reducing flowcell granularity to obtain high link utilization under small degrees of topology asymmetry. PDLB is only deployed on the sender without any modification on switch. We evaluate PDLB through large-scale NS2 simulations. The experimental results show that PDLB reduces the average flow completion time by up to ∼11-53% over the state-of-the-art load balancing schemes.

Performance Trade-Offs in Cyber–Physical Control Applications With Multi-Connectivity

Modern communication devices are often equipped with multiple wireless communication interfaces with diverse characteristics. This enables exploiting a form of multi-connectivity known as interface diversity to provide path diversity with multiple communication interfaces. Interface diversity helps to combat the problems suffered by single-interface systems due to error bursts in the link, which are a consequence of temporal correlation in the wireless channel. The length of an error burst is an essential performance indicator for cyber–physical control applications with periodic traffic, as this defines the period in which the control link is unavailable. However, the available interfaces must be correctly orchestrated to achieve an adequate trade-off between latency, reliability, and energy consumption. This work investigates how the packet error statistics from different interfaces impact the overall latency–reliability characteristics and explores mechanisms to derive adequate interface diversity policies. For this, we model the optimization problem as a partially observable Markov decision process (POMDP), where the state of each interface is determined by a Gilbert–Elliott model whose parameters are estimated based on experimental measurement traces from LTE and Wi-Fi. Our results show that the POMDP approach provides an all-round adaptable solution, whose performance is only 0.1% below the absolute upper bound, dictated by the optimal policy under the impractical assumption of full observability.

SRFabric: A Semi-Reconfigurable Rack Scale Topology

Rack scale design is a promising trend towards customized hardware design, where high density clusters of SoCs are integrated in the rack. One of the biggest challenges for rack scale computing is the interconnection network. Traditional data center topologies require too many ToR switches to support hundreds of SoCs, while distributed fabrics deliver a considerably high end-to-end latency and network oversubscription. Since no one topology fits all kinds of workloads, a flexible in-rack topology requires a careful redesign to dynamically adapt to diverse data center traffic within tight cost and space constraints in the rack. SRFabric is a semi-reconfigurable rack scale network topology that exploits the high path diversity, the cost-effectiveness of distributed fabrics, and the dynamic reconfigurability of circuit switches. This is accomplished by enabling multiple static ports and dynamic ports for each SoC. Leveraging the partial link reconfigurability, SRFabric is able to optimize its topology to dynamically adapt to various workload patterns. We further propose the design of SRFabric to decide the nearly optimal number of dynamic ports and static ports for expected communication density and performance. Extensive evaluations demonstrate that SRFabric can deliver lower average path length, i.e., 2.21 hops on average, and higher bisection bandwidth, i.e., up to 77% nonblocking bandwidth, and provide comparable performance with state-of-the-art strategy XFabric at a lower cost, i.e., XFabric costs up to 3 times more than that of SRFabric.

Distributed Reed-Muller Coded-Cooperative Spatial Modulation

This paper proposes the distributed Reed-Muller coded spatial modulation (DRMC-SM) scheme based on Kronecker product (KP) construction. This special construction enabled an effective distribution of classical Reed-Muller (RM) code along source and relay nodes. The proposed DRMC-SM scheme not only offers robustness in bit error rate (BER) performance but also enhances the spectral efficiency due to additional antenna index transmission inculcated by spatial modulation (SM). The usefulness of KP construction over classical Plotkin (CP) construction in coded-cooperation is analysed with and without incorporating SM. An efficient criteria for selecting the optimum bits is adopted at relay node which eventually results in better weight distribution of mutually constructed (source and relay) RM code under proposed KP construction. The numerical results show that proposed KP construction outperforms CP construction by gain of 1 dB in signal to noise ratio (SNR) at bit error rate (BER) of 7 x 10 -7 . Moreover, the proposed DRMC-SM scheme outperforms its non-cooperative Reed-Muller coded spatial modulation (RMC-SM) scheme as well as distributed turbo coded spatial modulation (DTC-SM) scheme in similar conditions. This prominent gain in SNR is evident due to path diversity, efficient selection of bits at relay node and the joint soft-in-soft-out (SISO) RM decoder deployed at the destination node.