Wednesday, 13 December 2017

Session 7: Performance Modelling and Prediction I

Enhancing Performance Prediction Robustness by Combining Analytical Modeling and Machine Learning

Authors:

Diego Didona (Universidade de Lisboa)
Francesco Quaglia (Sapienza Università di Roma)
Paolo Romano (Universidade de Lisboa)
Ennio Torre (Sapienza Università di Roma)

Abstract:

Classical approaches to performance prediction rely on two, typically antithetic, techniques: Machine Learning (ML) and Analytical Modeling (AM). ML takes a black box approach, whose accuracy strongly depends on the representativeness of the dataset used during the initial training phase. Specifically, it can achieve very good accuracy in areas of the features’ space that have been sufficiently explored during the training process. Conversely, AM techniques require no or minimal training, hence exhibiting the potential for supporting prompt instantiation of the performance model of the target system. However, in order to ensure their tractability, they typically rely on a set of simplifying assumptions. Consequently, AM’s accuracy can be seriously challenged in scenarios (e.g., workload conditions) in which such assumptions are not matched. In this paper we explore several hybrid/gray box techniques that exploit AM and ML in synergy in order to get the best of the two worlds. We evaluate the proposed techniques in case studies targeting two complex and widely adopted middleware systems: a NoSQL distributed key-value store and a Total Order Broadcast (TOB) service.

DOI: 10.1145/2668930.2688047

Full text: PDF

[#][]

A Comprehensive Analytical Performance Model of DRAM Caches

Authors:

Nagendra Gulur (Texas Instruments)
Mahesh Mehendale (Texas Instruments)
Ramaswamy Govindarajan (Indian Institute of Science)

Abstract:

Stacked DRAMpromises to offer unprecedented capacity, and bandwidth to multi-core processors at moderately lower latency than off-chip DRAMs. A typical use of this abundant DRAM is as a large last level cache. Prior research works are divided on how to organize this cache and the proposed organizations fall into one of two categories: (i) as a Tags-In-DRAM organization with the cache organized as small blocks (typically 64B) and metadata (tags, valid, dirty, recency and coherence bits) stored in DRAM, and (ii) as a Tags-In-SRAM organization with the cache organized as larger blocks (typiclly 512B or larger) and metadata stored on SRAM. Tags-In-DRAM organizations tend to incur higher latency but conserve off-chip bandwidth while the Tags-In-SRAM organizations incur lower latency at some additional bandwidth. In this work, we develop a unified performance model of the DRAM-Cache that models these different organizational styles. The model is validated against detailed architecture simulations and shown to have latency estimation errors of 10.7% and 8.8% on average in 4-core and 8-core processors respectively. We also explore two insights from the model: (i) the need for achieving very high hit rates in the metadata cache/predictor (commonly employed in the Tags-In-DRAM designs) in reducing latency, and (ii) opportunities for reducing latency by load-balancing the DRAM Cache and main memory.

DOI: 10.1145/2668930.2688044

Full text: PDF

[#][]

Systematically Deriving Quality Metrics for Cloud Computing Systems

Authors:

Matthias Becker (University of Paderborn)
Sebastian Lehrig (Chemnitz University of Technology)
Steffen Becker (Chemnitz University of Technology)

Abstract:

In cloud computing, software architects develop systems for virtually unlimited resources that cloud providers account on a pay-per-use basis. Elasticity management systems provision these resources autonomously to deal with changing workload. Such changing workloads call for new objective metrics allowing architects to quantify quality properties like scalability, elasticity, and efficiency, e.g., for requirements/SLO engineering and software design analysis. In literature, initial metrics for these properties have been proposed. However, current metrics lack a systematic derivation and assume knowledge of implementation details like resource handling. Therefore, these metrics are inapplicable where such knowledge is unavailable. To cope with these lacks, this short paper derives metrics for scalability, elasticity, and efficiency properties of cloud computing systems using the goal question metric (GQM) method. Our derivation uses a running example that outlines characteristics of cloud computing systems. Eventually, this example allows us to set up a systematic GQM plan and to derive an initial set of six new metrics. We particularly show that our GQM plan allows to classify existing metrics.

DOI: 10.1145/2668930.2688043

Full text: PDF

[#][]