Leonardo P. Tizzei

Job placement advisor based on turnaround predictions for HPC hybrid clouds

Several companies and research institutes are moving their CPU-intensive applications to hybrid High Performance Computing (HPC) cloud environments. Such a shift depends on the creation of software systems that help users decide where a job should be placed considering execution time and queue wait time to access on-premise clusters. Relying blindly on turnaround prediction techniques will affect negatively response times inside HPC cloud environments. This paper introduces a tool to make job placement decisions in HPC hybrid cloud environments taking into account the inaccuracy of execution and waiting time predictions. We used job traces from real supercomputing centers to run our experiments, and compared the performance between environments using real speedup curves. We also extended a state-of-the-art machine learning based predictor to work with data from the cluster scheduler. Our main findings are: (i) depending on workload characteristics, there is a turning point where predictions should be disregarded in favor of a more conservative decision to minimize job turnaround times and (ii) scheduler data plays a key role in improving predictions generated with machine learning using job trace data---our experiments showed around 20% prediction accuracy improvements.

Using Microservices and Software Product Line Engineering to Support Reuse of Evolving Multi-tenant SaaS

In order to achieve economies of scale, a Software as a Service (SaaS) should be configurable, multi-tenant efficient, and scalable. But building SaaS with these characteristics comes at a price of having more complex services. Some works in the literature integrate software product line engineering and service-oriented architecture to tackle the complexity of building multi-tenant SaaS. Most of these works focused on centralized approaches that rely on middleware or platforms, but they do not investigate the use of decentralized architectural style. Microservices architecture is an architectural style that relies on small, decentralized, and autonomous services that work together. Thus, this paper investigates the integrated use of microservices architecture and software produt line techniques to develop multi-tenant SaaS. We conducted an empirical study that analyzes the behavior of software reuse during the evolution of a multi-tenant SaaS. This empirical study showed an average software reuse of 62% of lines of code among tenants. We also provide lessons we learned during the the re-engineering and maintenance of such multi-tenant SaaS.

Evolving a Software Products Line for E-commerce Systems: a Case Study

Software Product Lines engineering is a technique that explores systematic reuse of software artifacts in large scale to implement applications that share a common domain and have some customized features. For improving Product Line Architecture evolution, it is advisable to develop Software Product Lines using a modular structure. This demand can be satisfied by an aspect-oriented and component-based feature-architecture method that integrates components, aspects and variation point aspect-connectors. This approach allows minimization of feature scattering in the architectural model and supports modular modelling of crosscutting features. A case study mapping major features of significant e-commerce systems operating in Brazil and other countries was performed to evaluate this approach. The assessment of our solution was performed comparing its stability and modularity with other two approaches. Our results indicate that change impact in the architectural model is reduced when using our solution in the context of Software Product Lines evolution.

WISE-SPL: bringing multi-tenancy to the weather InSights environment system

Weather conditions affect many cities and companies. The WISE (Weather In Sights Environment) system serves as a central place to gather and present weather related information for decision makers. It was initially developed to fit a single tenant. Due to a multi-tenant opportunity, WISE is evolving to be deployed on a Cloud environment to support on-demand computing resources and multiple clients. Software product line techniques were applied to model common and variable features of tenants. WISE-SPL enables the derivation of products for each client and also the deployment on Cloud infrastructure. The contribution of this work is a demonstration and discussion of benefits and limitations in applying SPL techniques, following a extractive approach, to build a multi-tenant Cloud application.

IBM Design Thinking Software Development Framework

The importance of understanding end user needs and involving them in the software development process is well known in software engineering. Agile Software Development methodologies have incorporated user feedback in different ways. User stories should represent the needs of a user, but often express the views of the Product Owner or the software development team. Several works have investigated integrating User Centered Design into Agile Software Development to satisfy end user needs. This work proposes a different approach focused on satisfying end user needs employing Design Thinking iterative software development. This methodology was applied in five real software development projects which have been analyzed as part of this work.

An SLA-Based Advisor for Placement of HPC Jobs on Hybrid Clouds

Several scientific and industry applications require High Performance Computing (HPC) resources to process and/or simulate complex models. Not long ago, companies, research institutes, and universities used to acquire and maintain on-premise computer clusters; but, recently, cloud computing has emerged as an alternative for a subset of HPC applications. This poses a challenge to end-users, who have to decide where to run their jobs: on local clusters or burst to a remote cloud service provider. While current research on HPC cloud has focused on comparing performance of on-premise clusters against cloud resources, we build on top of existing efforts and introduce an advisory service to help users make this decision considering the trade-offs of resource costs, performance, and availability on hybrid clouds. We evaluated our service using a real test-bed with a seismic processing application based on Full Waveform Inversion; a technique used by geophysicists in the oil & gas industry and earthquake prediction. We also discuss how the advisor can be used for other applications and highlight the main lessons learned constructing this service to reduce costs and turnaround times.

Architecting cloud tools using software product line techniques: an exploratory study

Multitenant cloud computing tools are usually complex and have to manage variabilities to support customization. Software Product Line (SPL) techniques have been successfully applied in the industry to manage variability in complex systems. However, few works in the literature discuss the application of SPL techniques to architect industry cloud computing tools, resulting in a lack of support to cloud architects on how to apply such techniques. This work presents how software product line techniques can be applied for architecting cloud tools, and discusses the benefits, drawbacks, and some challenges of applying such techniques to develop a real industry cloud tool, named as Installation Service.

Optimizing Multi-tenant Cloud Resource Pools via Allocation of Reusable Time Slots

Typical pricing models for IaaS cloud providers are slotted, using hour and month as time units for metering and charging resource usage. Such models lead to financial loss as applications may release resources much earlier than the end of the last allocated time slot, leaving the cost paid for the rest of the time unit wasted. This problem can be minimized for multi-tenant environments by managing resources as pools. This scenario is particularly interesting for universities and companies with various departments and SaaS providers with multiple clients. In this paper we introduce a tool that creates and manages resource pools for multi-tenant environments. Its benefit is the reduction of resource waste by reusing already allocated resources available in the pool. We discuss the architecture of this tool and demonstrate its effectiveness, using a seven-month workload trace obtained from a real multi-tenant SaaS financial risk analysis application. From our experiments, such tool reduced resource costs per day by 13 % on average in comparison to direct allocation of cloud provider resources.

A Self-tuning Scientific Framework using Model-Driven Engineering for Heterogeneous Execution Platforms

This article presents an ongoing work towards the extension of Sm@rtConfig -- a dynamic scheduling tool with self-tuning load-balancing functionalities targeting CPUs, GPUs, and other co-processors. This extension is based on the introduction of a high-level modeling phase for scientific applications. These applications are commonly complex, use (heterogeneous) high performance execution platforms, and require stakeholders of several disciplines. This way, it is important to raise abstraction level in earlier stages of development in order to deal with such complexities in an efficient way. By using Model-Driven Engineering, we propose an approach to transform Sm@rtConfig into a scientific framework comprising requirements engineering up to code generation for the target Processing Unit in which a task is scheduled at runtime. We advocate that our envisioned methodology facilitates not just cross-stakeholders development, but also replication of experimentations by the research community.