Programmable HSA Accelerators for Zynq UltraScale+ MPSoC Systems

Bauer, Wolfgang; Holzinger, Philipp; Reichenbach, Marc; Vaas, Steffen; Hartke, Paul; Fey, Dietmar

doi:10.1007/978-3-030-10549-5_57

Cited by 3 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The database community has long demonstrated the potential of "smart storage" [4,16,20] and now it is becoming mainstream. Amazon Aqua [1] provides query offloading to distributed storage using specialized hardware and there are commercially available Samsung SmartSSDs [36] appearing in the cloud as well.…”

Section: Emerging Smart Storage Devicesmentioning

confidence: 99%

“…As several studies showed, building smart storage only relying on small CPU cores (often by borrowing cores from the controller of the flash device itself) is very sensitive to the types of offloaded processing and can often become a severe bottleneck, instead of alleviating one [19,23]. As a result, the architecture of smart storage nodes typically incorporates some form of specialized ASIC [1,19] or a more general purpose Field Programmable Gate Array (FPGA) [4,16,20,36]. Figure 2 depicts a high level overview of a modern smart storage node, attached to fast networking to be used in disaggregated architectures.…”

Section: Emerging Smart Storage Devicesmentioning

confidence: 99%

See 1 more Smart Citation

Software-defined data protection

2021

View full text Add to dashboard Cite

Most modern data processing pipelines run on top of a distributed storage layer, and securing the whole system, and the storage layer in particular, against accidental or malicious misuse is crucial to ensuring compliance to rules and regulations. Enforcing data protection and privacy rules, however, stands at odds with the requirement to achieve higher and higher access bandwidths and processing rates in large data processing pipelines. In this work we describe our proposal for the path forward that reconciles the two goals. We call our approach "Software-Defined Data Protection" (SDP). Its premise is simple, yet powerful: decoupling often changing policies from request-level enforcement allows distributed smart storage nodes to implement the latter at line-rate. Existing and future data protection frameworks can be translated to the same hardware interface which allows storage nodes to offload enforcement efficiently both for company-specific rules and regulations, such as GDPR or CCPA. While SDP is a promising approach, there are several remaining challenges to making this vision reality. As we explain in the paper, overcoming these will require collaboration across several domains, including security, databases and specialized hardware design.

show abstract

Section: Emerging Smart Storage Devicesmentioning

confidence: 99%

Section: Emerging Smart Storage Devicesmentioning

confidence: 99%

Software-defined data protection

2021

View full text Add to dashboard Cite

show abstract

“…Two kinds of smart storage devices are developed recently: those which incorporate small CPU cores, such as ARM cores [16,20], and those which deploy specialized compute elements or FPGAs [17,13,33,3]. The former category offers more flexibility, but as explored in depth in the work of Koo et al [20], it can be difficult to predict the performance of the in-storage computation, and this can lead to performance degradation.…”

Section: Emerging Smart Storage Devicesmentioning

confidence: 99%

“…Overall, the SDP approach is beneficial in both cases because it reduces the complexity of the code running inside the storage nodes and allows developers to think of the enforcement as pipeline stages. We sketch our proposal targeting hardware platforms with a combination of ARM cored and FPGA fabric (as depicted in Figure 1), such as the Fidus Sidewinder-100 [3] or the Samsung SmartSSD [33,7]. The presence of both a general-purpose, albeit low power, CPU core, and a reprogrammable specialized hardware element ensures that the devices can be managed easily in a cloud setting and, at the same time, can deliver high-performance behavior for privacy-related processing.…”

Section: Emerging Smart Storage Devicesmentioning

confidence: 99%

Towards Software-Defined Data Protection: GDPR Compliance at the Storage Layer is Within Reach

István¹,

Ponnapalli²,

Chidambaram³

2020

Preprint

View full text Add to dashboard Cite

Enforcing data protection and privacy rules within large data processing applications is becoming increasingly important, especially in the light of GDPR and similar regulatory frameworks. Most modern data processing happens on top of a distributed storage layer, and securing this layer against accidental or malicious misuse is crucial to ensuring global privacy guarantees. However, the performance overhead and the additional complexity for this is often assumed to be significant -in this work we describe a path forward that tackles both challenges. We propose "Software-Defined Data Protection" (SDP), an adoption of the "Software-Defined Storage" approach to non-performance aspects: a trusted controller translates company and application-specific policies to a set of rules deployed on the storage nodes. These, in turn, apply the rules at line-rate but do not take any decisions on their own. Such an approach decouples often changing policies from request-level enforcement and allows storage nodes to implement the latter more efficiently.Even though in-storage processing brings challenges, mainly because it can jeopardize line-rate processing, we argue that today's Smart Storage solutions can already implement the required functionality, thanks to the separation of concerns introduced by SDP. We highlight the challenges that remain, especially that of trusting the storage nodes. These need to be tackled before we can reach widespread adoption in cloud environments.

show abstract

“…The traditional dynamic voltage frequency scaling (DVFS) [2] and adaptive voltage scaling (AVS) [3] techniques are focused on the collaborative architecture of an ARM processing system (PS) [4] and field programmable gate array (FPGA) programmable logic (PL) [5], and usually depend on the feedback controls. Furthermore, the optimum voltage and frequency can also be obtained by some modelling methods on the system power and performance, such as the static and dynamic synchronization-aware method [6], and nonlinear dynamics control modelling [7].…”

Section: Introductionmentioning

confidence: 99%

Energy and Performance Trade-Off Optimization in Heterogeneous Computing via Reinforcement Learning

Machado

Zahid

et al. 2020

Electronics

View full text Add to dashboard Cite

This paper suggests an optimisation approach in heterogeneous computing systems to balance energy power consumption and efficiency. The work proposes a power measurement utility for a reinforcement learning (PMU-RL) algorithm to dynamically adjust the resource utilisation of heterogeneous platforms in order to minimise power consumption. A reinforcement learning (RL) technique is applied to analyse and optimise the resource utilisation of field programmable gate array (FPGA) control state capabilities, which is built for a simulation environment with a Xilinx ZYNQ multi-processor systems-on-chip (MPSoC) board. In this study, the balance operation mode for improving power consumption and performance is established to dynamically change the programmable logic (PL) end work state. It is based on an RL algorithm that can quickly discover the optimization effect of PL on different workloads to improve energy efficiency. The results demonstrate a substantial reduction of 18% in energy consumption without affecting the application’s performance. Thus, the proposed PMU-RL technique has the potential to be considered for other heterogeneous computing platforms.

show abstract

Programmable HSA Accelerators for Zynq UltraScale+ MPSoC Systems

Cited by 3 publications

References 11 publications

Software-defined data protection

Software-defined data protection

Towards Software-Defined Data Protection: GDPR Compliance at the Storage Layer is Within Reach

Energy and Performance Trade-Off Optimization in Heterogeneous Computing via Reinforcement Learning

Contact Info

Product

Resources

About