Dynamic Resource Partitioning Research Articles

OPTIMUS: A Security-Centric Dynamic Hardware Partitioning Scheme for Processors that Prevent Microarchitecture State Attacks

Hardware virtualization allows multiple security-critical and ordinary (insecure) processes to co-execute on a processor. These processes temporally share hardware resources and endure numerous security threats on the microarchitecture state. State-of-the-art secure processor architectures, such as MI6 and IRONHIDE enable capabilities to execute security-critical processes in hardware isolated enclaves utilizing the strong isolation security primitive. The MI6 processor purges small state resources on each enclave entry/exit and statically partitions the last-level cache and DRAM regions to ensure strong isolation. IRONHIDE takes a spatial approach and creates two isolated clusters of cores in a multicore processor to ensure strong isolation for processes executing in the enclave cluster. Both architectures observe performance degradation due to static partitioning of shared hardware resources. OPTIMUS proposes a security-centric dynamic hardware resource partitioning scheme that operates entirely at runtime and ensures strong isolation. It enables deterministic resource allocations at the application level granularity, and limits the number of hardware reconfigurations to ensure bounded information leakage via the timing and termination channels. The dynamic hardware resource partitioning capability of OPTIMUS is shown to co-optimize performance and security for the MI6 and IRONHIDE architectures.

SDRP: Safe, Efficient, and SLO-Aware Workload Consolidation Through Secure and Dynamic Resource Partitioning

Workload consolidation is a widely-used technique to improve the resource utilization of services computing systems by consolidating latency-critical (LC) and batch workloads on the same physical server. The resource manager for workload consolidation dynamically allocates hardware resources (e.g., cores, caches) to the workloads to maximize the resource utilization while satisfying the service-level objective (SLO) of the LC workloads. Since security-critical hardware resources are dynamically allocated across consolidated workloads, information leakages can be created among workloads through microarchitectural side-channel (SC) attacks. Despite extensive prior works, it is yet to investigate efficient system software support for achieving high resource utilization without compromising the SLO and security of consolidated workloads. To bridge this gap, we propose SDRP, secure and dynamic resource partitioning for safe, efficient, and SLO-aware workload consolidation. As with the state-of-the-art techniques, SDRP dynamically allocates hardware resources to enhance the resource utilization and provide the SLO guarantees. In contrast to the state-of-the-art techniques, SDRP dynamically sanitizes security-critical hardware resources to robustly defeat microarchitectural SC attacks. Our quantitative evaluation demonstrates that SDRP achieves high resource sanitization quality, introduces low performance overheads, delivers high resource utilization with the SLO and security guarantees, and defeats the last-level cache (LLC)-based SC attack.

Evidence‐based dynamic radio resource allocation to mitigate inter cell interference employing cooperative communication

The paper proposes a framework for traffic based dynamic radio resource allocation to mitigate Inter Cell Interference (ICI) in cooperative communication for LTE MIMO-OFDMA systems using evidence theory. The objective is to address ICI, by dynamic user categorization, resource partitioning and management of radio resources. Through evidence based dynamic categorization of users, dynamic frequency planning and dynamic allocation of resources, spectral efficiency is enhanced by increasing data rate of users. It requires the cell users to be categorized into edge and center users depending on the confidence factor, computed using Dempster Shaffer Combinational Rule (DSCR). A traffic based dynamic Soft Frequency Reuse (SFR) plan is proposed using categorized user information to partition the available resources i.e., subcarrier and associated power. A subset of resources is allocated to identified edge users based on dynamic traffic. The traffic based evidential categorization of users, partitioning of resources provides improved data rate and spectral efficiency of edge users. The allocation of resources is dynamic in the proposed framework, leading to an improvement of spectral efficiency from 6.6 b/s/Hz to 12.4 b/s/Hz in a cell. The proposed methodology adapts to the user traffic in a cell and allocates the resources satisfying maximum users.

Dynamic competition and resource partitioning during the early life of two widespread, abundant and ecologically similar fishes

Competition and resource partitioning can have profound implications for individuals, populations and communities, and thus food webs, ecosystems and the management of biota and environments. In many species, the impacts of competition and resource partitioning are believed to be most severe during early life, but our understanding of the mechanisms and implications is incomplete. This study revealed short-term variations in both the occurrence and direction of competition during the early life of roach Rutilus rutilus and common bream Abramis brama, two of the most widespread and abundant fish species in Europe. There was also evidence of resource partitioning when small taxa dominated the zooplankton, but not when larger taxa were more abundant. In spite of the differences in foraging ecology, there were no significant differences in growth or nutritional condition in allopatry and sympatry. Similar to the concept of condition-specific competition, when competitive abilities vary along environmental gradients, the impacts of interspecific interactions on foraging ecology, growth and condition are dynamic and likely vary according to temporal fluctuations in prey availability. This is important because short-term incidences of competition could have cascading effects on food webs, even when no impacts on growth rates or condition are detected.

Evidence for dynamic resource partitioning between two sympatric reef shark species within the British Indian Ocean Territory.

Stable‐isotope analyses (δ13C, δ15N and δ34S) of multiple tissues (fin, muscle, red blood cells and plasma), revealed ontogenetic shifts in resource use by grey reef sharks Carcharhinus amblyrhynchos and resource partitioning with silvertip sharks Carcharhinus albimarginatus within the British Indian Ocean Territory marine protected area (MPA). Resource partitioning varied temporally, with C. albimarginatus feeding on more pelagic prey during October to January, potentially attributable to an influx of pelagic prey from outside the MPA at that time. Reef sharks may therefore be affected by processes outside an MPA, even if the sharks do not leave the MPA.

An integrative circuit-host modelling framework for predicting synthetic gene network behaviours.

One fundamental challenge in synthetic biology is the lack of quantitative tools that accurately describe and predict the behaviours of engineered gene circuits. This challenge arises from multiple factors, among which the complex interdependence of circuits and their host is a leading cause. Here we present a gene circuit modelling framework that explicitly integrates circuit behaviours with host physiology through bidirectional circuit-host coupling. The framework consists of a coarse-grained but mechanistic description of host physiology that involves dynamic resource partitioning, multilayered circuit-host coupling including both generic and system-specific interactions, and a detailed kinetic module of exogenous circuits. We showed that, following training, the framework was able to capture and predict a large set of experimental data concerning the host and its foreign gene overexpression. To demonstrate its utility, we applied the framework to examine a growth-modulating feedback circuit whose dynamics is qualitatively altered by circuit-host interactions. Using an extended version of the framework, we further systematically revealed the behaviours of a toggle switch across scales from single-cell dynamics to population structure and to spatial ecology. This work advances our quantitative understanding of gene circuit behaviours and also benefits the rational design of synthetic gene networks.

Joint Downlink User Association and Interference Management in Two-Tier HetNets With Dynamic Resource Partitioning

We investigate a joint user association and interference management problem in two-tier downlink heterogeneous networks where pico base stations (BSs) are densely deployed in areas of high traffic demand. We employ macro almost blanking subframes (ABSs) to mitigate cross-tier interference. To manage co-tier interference among picocells, we introduce pico operation mode (POM): During each POM, a distinct subset of picocells serve users simultaneously; different POMs operate at different times at different durations. We formulate the problem as maximizing the network utility under proportional fairness with the optimization over user association and resource partitioning (RP) (including the macro ABS and the amount of time allocated to each POM). As an exhaustive search over all possible POMs in the optimization may become computationally infeasible, we propose a method of preselecting favorable POMs to reduce the dimension of the optimization. With the selected POMs, we explicitly examine the structure of the optimal solutions and show that 1) the optimal user association is load-aware and can be determined by rate bias values of each BS, and 2) all the POMs and the macro BSs have balanced load in the sense that the ratios of the number of associated users to the allocated time are the same. Based on the analysis, an alternating algorithm is developed to obtain the RP and the biases. We demonstrate through numerical examples that 1) the proposed POM selection method does not incur significant performance loss, compared with the case where all possible POMs are considered; 2) the alternating algorithm provides near-optimal cell association and RP solutions; 3) by applying the proposed framework, significant network performance improvement can be achieved with dense pico deployments, compared with baselines without co-tier interference management and baselines with sparse pico deployments.

Dynamic Resource Partitioning for Heterogeneous Multi-Core-Based Cloud Computing in Smart Cities

As the smart cities emerged for more comfortable urban spaces, services, such as health, transportation, and so on, need to be promoted. In addition, the cloud computing provides flexible allocation, migration of services, and better security isolation; therefore, it is the infrastructure for the smart cities. Single instruction-set architecture (ISA) heterogeneous multi-core processors have higher performance per watt than their symmetric counterparts and are popular in current processors. In current cloud computing, which integrates a few fast out-of-order cores, coupled with a large number of simpler, slow cores, all cores expose the same ISA. The best way to leverage the effectiveness of these systems is to accelerate sequential CPU-bound threads using fast cores, and to improve the throughput of parallel memory-bound threads using slow cores. However, shared hardware resources, such as memory, respond to requests from all cores, which interfere with each other, leading to both low speed for fast cores and low throughput for slow cores. In this paper, we propose a dynamic resource partitioning (DRP) method for single-ISA heterogeneous multi-cores, which partitions the shared resources according to both threads’ requirements for the shared resources and the performance of their running cores. The key principle is to profile both threads’ resource characteristics at run-time and the performance of the cores that the threads are running on to estimate demands for resources. Then, we use the estimation to direct our resource partitioning. Moreover, we integrate our DRP with current memory scheduling policies to improve the system performance further for the two methods being orthogonal.

Emergence of a dynamic resource partitioning based on the coevolution of phenotypic plasticity in sympatric species

This paper investigates the coevolutionary dynamics of the phenotypic plasticity in the context of overlap avoidance behaviors of shared niches in sympatric species. Especially, we consider whether and how a differentiation of phenotypic plasticity can emerge under the assumption that there are no initial asymmetric relationships among coevolving species. We construct a minimal model where several different species participate in a partitioning of their shared niches, and evolve their behavioral plasticity to avoid an overlap of their niche use. By conducting evolutionary experiments with various conditions of the number of species and niches, we show that the two different types of asymmetric distributions of phenotypic plasticity emerge depending on the settings of the degree of congestion of the shared niches. In both cases, all species tended to obtain the similar amount of fitness regardless of such differences in their plasticity. We also show that the emerged distributions are coevolutionarily stable in general.

Designing and dynamically load balancing hybrid LU for multi/many-core

Designing high-performance LU factorization for modern hybrid multi/many-core systems requires highly-tuned BLAS subroutines, hiding communication latency and balancing the load across devices of variable processing capabilities. In this paper we show how single-precision LU factorization is accelerated on Intel® MIC(Many Integrated Core) architecture in both native and hybrid (Intel® Xeon® processor and Intel MIC) configurations. Our SGEMM implementation delivers close to 1 Tflop/s on Intel's first implementation of Intel MIC architecture [codenamed Knight's Ferry (KNF)] silicon platform. Our implementation takes full advantage of multiple levels of memory hierarchy on MIC, and successfully utilizes up to 80% of its peak compute capability. Our LU factorization performance exceeds 570 Gflop/s including matrix transfer overhead when executed entirely on a KNF coprocessor. Our hybrid implementation, which offloads parts of LU processing to a dual-socket multi-core Intel Xeon processor X5680 host, delivers up to 772 Gflop/s. The novel aspect of our implementations is dynamic resource partitioning to improve load balance across the entire system.

Pollinator diversity and specialization in relation to flower diversity

In the face of global decline in biodiversity, the relationship between diversity and species interactions deserves particular attention. If pollinators are strongly dependent on floral diversity due to mutual specialization, declines in plant diversity, e.g. caused by land use intensification, may be associated with linked extinctions of pollinators. However, the general extent of pollinator specialization is still poorly known. To explore the dependence of local bee and hoverfly communities on flower diversity, we recorded flower supply and flower-visiting insects on 27 meadows with varying flower diversity in southern Germany and analyzed (a) whether the diversity of flower visitors is correlated with flower diversity, (b) whether the degree of dietary specialization of flower visitors changes with flower diversity and (c) whether flower preferences of individual flower visitor species are constant or variable between different communities. Flower-visitor interaction webs were compiled during a single day on each meadow. This approach prevents relating pollinator species to flowers they never encounter because of non-overlapping phenology or spatial segregation. (a) Flower diversity and flower visitor diversity were positively correlated. (b) Flower visitor assemblies were significantly specialized at a relatively high level, contrasting to the opinion that plant-pollinator webs are highly generalized, and providing a possible explanation for the positive diversity correlation. However, the level of specialization did not change significantly across the gradient of flower diversity, suggesting that pollinators are partitioned to a similar extent in each meadow. (c) In the analysis of ten common flower visitor species previously categorized as generalists, strong evidence was found for both, consistent preferences and preferences that differ between sites. These results indicate a flexibility in flower preferences and a dynamic resource partitioning among pollinators. Generally, our findings highlight the complexity of plant-pollinator interactions and confirm the importance of flower diversity for bee and hoverfly communities.

Dynamic Resource Partitioning for Downlink Femto-to-Macro-Cell Interference Avoidance

Femto-cells consist of user-deployed Home Evolved NodeBs (HeNBs) that promise substantial gains in system spectral efficiency, coverage, and data rates due to an enhanced reuse of radio resources. However, reusing radio resources in an uncoordinated, random fashion introduces potentially destructive interference to the system, both, in the femto and macro layers. An especially critical scenario is a closed-access femto-cell, cochannel deployed with a macro-cell, which imposes strong downlink interference to nearby macro user equipments (UEs) that are not permitted to hand over to the femto-cell. In order to maintain reliable service of macro-cells, it is imperative to mitigate the destructive femto-cell to macro-cell interference. The contribution in this paper focuses on mitigating downlink femto-cell to macro-cell interference through dynamic resource partitioning, in the way that HeNBs are denied access to downlink resources that are assigned to macro UEs in their vicinity. By doing so, interference to the most vulnerable macro UEs is effectively controlled at the expense of a modest degradation in femto-cell capacity. The necessary signaling is conveyed through downlink high interference indicator (DL-HII) messages over the wired backbone. Extensive system level simulations demonstrate that by using resource partitioning, for a sacrifice of 4% of overall femto downlink capacity, macro UEs exposed to high HeNB interference experience a tenfold boost in capacity.