{"id":770,"date":"2023-08-09T14:03:38","date_gmt":"2023-08-09T14:03:38","guid":{"rendered":"https:\/\/cloud.ic.uff.br\/?page_id=770"},"modified":"2023-09-16T18:37:16","modified_gmt":"2023-09-16T18:37:16","slug":"projeto-capes-print","status":"publish","type":"page","link":"https:\/\/cloud.ic.uff.br\/index.php\/pt\/projeto-capes-print\/","title":{"rendered":"CAPES PrInt"},"content":{"rendered":"\n
Ger\u00eancia de Recursos em Nuvens para Execu\u00e7\u00e3o de Aplica\u00e7\u00f5es de Alto Desempenho<\/strong><\/h5>\n\n\n\n

Coordenadora: L\u00facia Maria de Assump\u00e7\u00e3o Drummond<\/span><\/p>\n\n\n\n

O projeto visa resolver o problema de aloca\u00e7\u00e3o e ger\u00eancia de recursos em nuvens de computadores para aplica\u00e7\u00f5es HPC, minimizando o tempo de execu\u00e7\u00e3o, consumo de energia e maximizando a toler\u00e2ncia a falhas, sem violar acordos de n\u00edvel de servi\u00e7o (SLA), e usando aplica\u00e7\u00f5es da biologia como estudo de caso. Tradicionalmente a computa\u00e7\u00e3o em nuvem tem sido usada para compartilhamento de dados e servi\u00e7os de uso geral, entretanto esta tamb\u00e9m tem aparecido como uma alternativa promissora para executar aplica\u00e7\u00f5es de HPC (High Performance Computing), mais recentemente. Este paradigma computacional oferece diversas vantagens quando comparado a uma infraestrutura dedicada, tais como o r\u00e1pido provisionamento de recursos e significativa redu\u00e7\u00e3o de custos operacionais.<\/p>\n\n\n\n

Entretanto, alguns desafios devem ser superados para reduzir a diferen\u00e7a entre o desempenho oferecido por uma infraestrutura dedicada e pelas nuvens.  Overheads introduzidos pela camada de virtualiza\u00e7\u00e3o, heterogeneidade de hardware e altas lat\u00eancias de rede afetam negativamente o desempenho da aplica\u00e7\u00f5es HPC. Al\u00e9m disso, provedores de nuvens, geralmente adotam pol\u00edticas de compartilhamento de recursos que podem reduzir ainda mais o desempenho de tais aplica\u00e7\u00f5es. Tipicamente, um servidor f\u00edsico pode hospedar v\u00e1rias m\u00e1quinas virtuais que podem causar conten\u00e7\u00e3o no acesso a recursos compartilhados, como cache e mem\u00f3ria principal, reduzindo significativamente seus desempenhos.<\/p>\n\n\n\n

Al\u00e9m disso, a sele\u00e7\u00e3o de m\u00e1quinas virtuais e sua configura\u00e7\u00e3o manual \u00e9 tarefa bastante complexa para cientistas que desenvolvem aplica\u00e7\u00f5es HPC e n\u00e3o s\u00e3o experts em ferramentas de administra\u00e7\u00e3o de nuvens. Os escalonadores de aplica\u00e7\u00f5es que possuem diversas pol\u00edticas que variam de acordo com a fun\u00e7\u00e3o objetivo como minimizar o tempo de execu\u00e7\u00e3o total, minimizar a demanda por energia, mantendo uma garantia de n\u00edvel de servi\u00e7o com o usu\u00e1rio, entre outras, tem papel fundamental para garantir a efici\u00eancia da execu\u00e7\u00e3o de tais aplica\u00e7\u00f5es. A fim de alavancar o uso de nuvens para execu\u00e7\u00e3o de aplica\u00e7\u00f5es HPC, este projeto visa tratar esses diversos aspectos. A import\u00e2ncia do uso de nuvens para execu\u00e7\u00e3o de aplica\u00e7\u00f5es HPC, pode ser observada por algumas iniciativas, tal como o UberCloud, que tem oferecido servi\u00e7o de HPC na nuvem, onde os usu\u00e1rios podem discutir a experi\u00eancia de usar tal ambiente. Como estudo de caso consideramos principalmente experimentos na \u00e1rea de bioinform\u00e1tica e, em especial, a gen\u00f4mica comparativa.<\/p>\n\n\n\n

\n\n\n\n
Equipe<\/strong><\/h5>\n\n\n\n
Universidade Federal Fluminense<\/strong><\/th><\/tr><\/thead>
L\u00facia Maria de Assump\u00e7\u00e3o Drummond<\/span>   \"\"<\/a>   \"\"<\/a>   <\/td><\/tr>
Maria Cristina Silva Boeres<\/span>   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Eugene Francis Vinod Rebello<\/span>   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Yuri Abitbol de Menezes Frota   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Igor Monteiro Moraes   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Jos\u00e9 Viterbo Filho<\/span>   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Daniel Cardoso Moraes de Oliveira   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Eduardo Uchoa   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Artur Pessoa   \"\"<\/a>   \"\"<\/a><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Sorbonne Universit\u00e9<\/strong><\/th><\/tr><\/thead>
Pierre Sens<\/span>   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Luciana Arantes<\/span>   \"\"<\/a>   \"\"<\/a><\/td><\/tr>
Guy Pujolle<\/span>   \"\"<\/a><\/td><\/tr>
Amal El Fallah Seghrouchni   \"\"<\/a>   \"\"<\/a><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Universit\u00e9 d\u2019Avignon<\/strong><\/th><\/tr><\/thead>
Rosa Figueiredo   \"\"<\/a>   \"\"<\/a><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Universit\u00e9 de Montpellier<\/strong><\/th><\/tr><\/thead>
Esther Pacitti   \"\"<\/a>   \"\"<\/a><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Universit\u00e9 de Bordeaux<\/strong><\/th><\/tr><\/thead>
Fran\u00e7ois Vanderbeck   \"\"<\/a><\/td><\/tr>
Ruslan Sadykov   \"\"<\/a><\/td><\/tr>
La\u00e9rcio Lima Pilla   \"\"<\/a>   \"\"<\/a>   \"\"<\/a><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
MINES ParisTech<\/strong><\/th><\/tr><\/thead>
Claude Tadonki<\/span>   \"\"<\/a>   \"\"<\/a><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\n\n\n\n
Editais<\/strong><\/h5>\n\n\n\n
Edital<\/strong><\/th>Links<\/strong><\/th>Per\u00edodo de Inscri\u00e7\u00e3o<\/strong><\/th>Situa\u00e7\u00e3o<\/strong><\/th><\/tr><\/thead>
Sele\u00e7\u00e3o de Bolsista PDSE 2019<\/td>Edital<\/a>
Anexo I<\/a>
Anexo II<\/a><\/td>		15\/03\/2019 a 01\/04\/2019<\/td>Encerrado<\/td><\/tr>
Professor Visitante no Exterior J\u00fanior<\/td>Edital<\/a><\/td>01\/07\/2019 a 08\/07\/2019<\/td>Encerrado<\/td><\/tr>
Pesquisador Visitante no Brasil<\/td>Edital<\/a><\/td>01\/07\/2019 a 15\/07\/2019<\/td>Encerrado<\/td><\/tr>
Sele\u00e7\u00e3o de Bolsista PDSE 2020<\/td>Edital<\/a>
Anexo I<\/a>
Anexo II<\/a>
Ata de Sele\u00e7\u00e3o<\/a><\/td>		13\/10\/2020 a 19\/10\/2020<\/td>Encerrado<\/td><\/tr>
Professor Visitante no Exterior S\u00eanior<\/td>Edital<\/a>
Ata de Sele\u00e7\u00e3o<\/a><\/td>		26\/10\/2020 a 02\/11\/2020<\/td>Encerrado<\/td><\/tr>
Pesquisador Visitante no Brasil<\/td>Edital<\/a><\/td>12\/11\/2020 a 13\/11\/2020<\/td>Encerrado<\/td><\/tr>
Pesquisador Visitante no Brasil<\/td>Edital<\/a><\/td>10\/08\/2022 a 19\/08\/2022<\/td>Encerrado<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\n\n\n\n
Eventos<\/strong><\/h5>\n\n\n\n
Evento<\/strong><\/th>Data<\/strong><\/th>Local<\/strong><\/th><\/tr><\/thead>
Palestra: Smart City Network Uberization<\/summary>

Abstract: Connectivity is a key issue for smart cities. Providing Internet to a large number of users and things requires intensive network densification, an operation consisting in increasing the number of access points in order to provide a high density of users and things with an Internet connection. This presentation describes how to reach this goal using an uberization solution. We compare our solution with 5G and Fog networking solutions. We also describe how it is possible to secure the network with blockchains solutions.<\/p><\/details><\/em>
Guy Pujolle<\/span> (Professor Visitante)<\/td>

26\/03\/2019
\u00e0s 11h<\/td>		Audit\u00f3rio do IC\/UFF<\/td><\/tr>
Defesa de Proposta de Tese de Doutorado: A Hibernation-Aware Dynamic Scheduler for Cloud Environments<\/summary>

Abstract: Cloud platforms have emerged as a prominent environment to execute different classes of applications providing on-demand resources as well as scalability. They usually offer several types of Virtual Machines (VMs) which have different guarantees in terms of availability and volatility, provisioning the same resource through multiple pricing models. For instance, in the Amazon EC2 cloud, the user pays per hour for on-demand VMs while spot VMs are unused instances available for a lower price. Despite the financial advantages, a spot VM can be terminated or hibernated by EC2 at any moment.

Using both hibernation-prone spot VMs (for cost sake) and on-demand VMs, we propose in this paper the Hibernation-Aware Dynamic Scheduler (HADS), that uses those VMs to execute applications composed of independent tasks (bag-of-task) with deadline constraints. Besides that, we also define the problem of temporal failures, that occurs when a spot VM hibernates, and it does not resume within a time that guarantees the application\u2019s deadline. Our scheduling approach, thus, aims at minimizing the monetary costs of bag-of-tasks applications in EC2 cloud, respecting its deadline even in the presence of hibernation, and avoiding temporal failures. Performance results with real executions using Amazon EC2 VMs confirm the effectiveness of our scheduling and that it can tolerate temporal failures.<\/p><\/details><\/em>
Luan Teylo<\/span> (Aluno de Doutorado)

Orientadores: L\u00facia Drummond<\/span>, Luciana Arantes<\/span> e Pierre Sens<\/span><\/td>

03\/05\/2019
\u00e0s 10h30<\/td>		IC Sala 317<\/td><\/tr>
Palestra: Dealing with Non Uniform Memory Access<\/summary>

Abstract: A natural trend in the multicore evolution is the increase in the number of integrated cores, while keeping the configuration of a shared main memory. One common packaging considers a Non Uniform Memory Access (NUMA), where the overall available memory is made of several blocks that are physically separated but interconnected so as to form a virtually contiguous memory with a unified addressing. Form the programming point of view, this aspect is completely virtual, and ordinary programmers are not even aware of this technical reality. The main consequence of not addressing the NUMA configuration is the unacceptable scalability that can be observed using a standard parallelization. This penalty is the conjunction of remote accesses and bus contention, among others. In this talk, we will explain the concept, how is it technically considered, what are the effects and how to deal with this programming concern.<\/p><\/details><\/em>
Claude Tadonki<\/span> (Professor Visitante)<\/td>

15\/05\/2019
\u00e0s 10h<\/td>		IC Sala 202<\/td><\/tr>
Palestra: Influence of Tasks Duration Variability on Task-Based Runtime Schedulers<\/summary>

Abstract: In the context of HPC platforms, individual nodes nowadays consist of heterogenous processing resources such as GPU units and multicores. Those resources share communication and storage resources, inducing complex co-scheduling effects, and making it hard to predict the exact duration of a task or of a communication. To cope with these issues, runtime dynamic schedulers such as StarPU have been developed. These systems base their decisions at runtime on the state of the platform and possibly on static priorities of tasks computed offline. In this paper, our goal is to quantify performance variability in the context of HPC heterogeneous nodes, by focusing on very regular dense linear algebra kernels, such as Cholesky and LU factorizations. We therefore first concentrate on the evaluation of the individual block-size kernels variability. Then, we analyze the impact of this variability at the scale of a full application on a dynamic runtime scheduler such as StarPU, in order to analyze whether the strategies that have been designed in the context of MapReduce applications to cope with stragglers could be transferred to HPC systems, or if the dynamic nature of runtime schedulers is enough to cope with actual performance variations, even in presence of task dependencies.<\/p><\/details><\/em>
Olivier Beaumont<\/span> (Professor Visitante)<\/td>

29\/05\/2019
\u00e0s 11h<\/td>		IC Sala 202<\/td><\/tr>
Palestra: Optical Interconnects for Advanced Proceessing Systems<\/summary>

Abstract: The presentation intends to identify the role of optical interconnects in next-generation distributed processing systems and will review the challenges and recent developments in board-level and on-board chip-to-chip interconnection for Data Centres and High-Performance Compute applications.<\/p><\/details><\/em>
George T. Kanellos<\/span> (Professor Visitante)<\/td>

29\/10\/2019
\u00e0s 15h<\/td>		Audit\u00f3rio do IC\/UFF<\/td><\/tr>
Palestra: Failure Detection in Large Distributed Systems<\/summary>

Abstract: Failure detection is a prerequisite to failure mitigation and a key component to build distributed algorithms requiring resilience. This talk introduces the problem of failure detection in asynchronous network where the transmission delay is not known. We show how distributed failure detector oracles can be used to address fundamental problems such as consensus, k-set agreement, or mutual exclusion. Finally, we focus on how to build scalable failure detectors.<\/p><\/details><\/em>
Pierre Sens<\/span> (Professor Visitante)<\/td>

29\/10\/2019
\u00e0s 16h<\/td>		Audit\u00f3rio do IC\/UFF<\/td><\/tr>
Palestra: A Communication-Efficient Causal Broadcast Protocol<\/summary>

Abstract: A causal broadcast ensures that messages are delivered to all nodes (processes) preserving causal relation of the messages. We have proposed [ICPP 2018] a causal broadcast protocol for distributed systems whose nodes are logically organized in a virtual hypercube like topology called Vcube. Messages are broadcast by dynamically building spanning trees rooted in the message\u2019s source node. By using multiple trees, the contention bottleneck problem of a single root spanning tree approach is avoided. Furthermore, different trees can intersect at some node. Hence, by taking advantage of both the out-of-order reception of causally related messages at a node and these paths intersections, a node can delay to one or more of its children in the tree, the forwarding of the messages whose some causal dependencies it knows that the children in question cannot satisfy yet. Such a delay does not induce any overhead. Experimental evaluation conducted on top of PeerSim simulator confirms the communication effectiveness of our causal broadcast protocol in terms of latency and message traffic reduction.<\/p><\/details><\/em>
Luciana Arantes<\/span> (Professor Visitante)<\/td>

30\/10\/2019
\u00e0s 11h<\/td>		Audit\u00f3rio do IC\/UFF<\/td><\/tr>
Palestra: Probabilistic Byzantine Tolerance Scheduling in Hybrid Cloud Environments<\/summary>

Abstract: This talk explores scheduling challenges in providing probabilistic Byzantine fault tolerance in a hybrid cloud environment, consisting of nodes with varying reliability levels, compute power, and monetary cost. In this context, the probabilistic Byzantine fault tolerance guarantee refers to the confidence level that the results of a given computation is correct despite potential Byzantine failures. We formally define a family of such scheduling problems distinguished by whether they insist on meeting a given latency limit and trying to optimize the monetary budget or vice versa. For the case where the latency bound is a restriction and the budget should be optimized, we present several heuristic protocols and compare between them using extensive simulations.<\/p><\/details><\/em>
Pierre Sens<\/span> (Professor Visitante)<\/td>

30\/10\/2019
\u00e0s 12h<\/td>		Audit\u00f3rio do IC\/UFF<\/td><\/tr>
Workshop: 1\u00ba WCloud-HPC<\/a><\/em>

L\u00facia Drummond<\/span> (Organizador)<\/td>		31\/08\/2020 e 01\/09\/2020<\/td>Online<\/td><\/tr>
Workshop: 2\u00ba WCloud-HPC<\/a><\/em>

L\u00facia Drummond<\/span> (Organizador)<\/td>		16\/09\/2021 e 17\/09\/2021<\/td>Online<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\n\n\n\n
Miss\u00f5es<\/strong><\/h5>\n\n\n\n
Miss\u00e3o<\/strong><\/th>Per\u00edodo<\/strong><\/th>Local<\/strong><\/th><\/tr><\/thead>
Palestra: A Hibernation Aware Scheduler for Cloud Environments<\/summary>

Abstract: Nowadays, cloud platforms usually offer several types of Virtual Machines (VMs) which have different guarantees in terms of availability and volatility, provisioning the same resource through multiple pricing models. For instance, in the Amazon EC2 cloud, the user pays per hour for on-demand VMs while spot VMs are unused instances available for a lower price. Despite the monetary advantages, a spot VM can be terminated or hibernated by EC2 at any moment.

In this talk, we present the Hibernation-Aware Dynamic Scheduler (HADS), to schedule applications composed of independent tasks (bag-of-tasks) with deadline constraints in both hibernation-prone spot VMs (for cost sake) and on-demand VMs. We also consider the problem of temporal failures, that occurs when a spot VM hibernates, and does not resume within a time that guarantees the application\u2019s deadline. Our dynamic scheduling approach aims at minimizing the monetary costs of bag-of-tasks applications execution, respecting its deadline even in the presence of hibernation. It is also able to avoid temporal failures, by using task migration and work-stealing techniques. Experimental results with real executions using Amazon EC2 VMs confirm the effectiveness of our scheduling when compared with on-demand VM only based approaches, in terms of monetary costs and execution times. It is also shown that our strategy can tolerate temporal failures..<\/p><\/details><\/em>
L\u00facia Drummond<\/span><\/td>

14\/11\/2019 a 25\/11\/2019<\/td>Sorbonne Universit\u00e9, Paris<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\n\n\n\n
Bolsas<\/strong><\/h5>\n\n\n\n
Bolsa<\/strong><\/th>Per\u00edodo<\/strong><\/th>Local<\/strong><\/th><\/tr><\/thead>
Bolsa de Doutorado Sanduiche
(Seis Meses)<\/summary>

Tese: A Dynamic Task Scheduler Tolerant to Multiple Hibernations in Cloud Environments.<\/p><\/details><\/em>
Luan Teylo<\/span><\/td>

09\/2019 a 02\/2020<\/td>Sorbonne Universit\u00e9, Paris<\/td><\/tr>
Bolsa de Pesquisador Visitante no Brasil<\/summary>

Palestra: A Communication-Efficient Causal Broadcast Protocol.<\/p><\/details><\/em>
Luciana Arantes<\/span><\/td>

11\/2019<\/td>Universidade Federal Fluminense, Niter\u00f3i<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\n\n\n\n
Projetos Aprovados<\/strong><\/h5>\n\n\n\n
Projeto<\/strong><\/th>Coordenador<\/strong><\/th><\/tr><\/thead>
STIC AMSUD \u2013 Edital n\u00ba 10\/2019<\/summary>

\n
Nome do Programa-Capes: STIC AMSUD \u2013 COOPERA\u00c7\u00c3O EM CI\u00caNCIA E TECNOLOGIA DA INFORMA\u00c7\u00c3O E DA COMUNICA\u00c7\u00c3O FRAN\u00c7A \u2013 AM\u00c9RICA DO SUL \u2013 CAPES\/CDEFI

Coordenador: EDUARDO UCHOA BARBOZA

N\u00famero do instrumento de sele\u00e7\u00e3o: STIC AMSUD \u2013 Edital no 10\/2019 \u2013 Projetos

Calend\u00e1rio
– O primeiro ano do projeto inicia em 01\/2020 e finaliza em 12\/2020;
– O segundo ano do projeto do projeto inicia em 01\/2021 e finaliza em 12\/2021.

Segundo o edital do programa STIC AMSUD \u2013 COOPERA\u00c7\u00c3O EM CI\u00caNCIA E TECNOLOGIA DA INFORMA\u00c7\u00c3O E DA COMUNICA\u00c7\u00c3O FRAN\u00c7A \u2013 AM\u00c9RICA DO SUL \u2013 CAPES\/CDEFI, a CAPES subsidiar\u00e1 o projeto e a equipe brasileira, conforme disponibilidade or\u00e7ament\u00e1ria anual, em: 4 (quatro) Miss\u00f5es de trabalho sendo at\u00e9 2 (duas) para a Fran\u00e7a e 2 (duas) para pa\u00edses da Am\u00e9rica do Sul, para o coordenador ou membros doutores da equipe inscritos no projeto. Cada miss\u00e3o de trabalho dever\u00e1 ter dura\u00e7\u00e3o m\u00ednima de 07 (sete) e m\u00e1xima de 20 (vinte) dias; At\u00e9 2 (duas) bolsas de estudos por ano, nas modalidades: Doutorado-Sandu\u00edche: dura\u00e7\u00e3o de 4 (quatro) a 12 (doze) meses; P\u00f3s-Doutorado: dura\u00e7\u00e3o de 2 (dois) a 12 (doze) meses.<\/p><\/details><\/td>

Eduardo Uchoa Barboza<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\n\n\n\n
Publica\u00e7\u00f5es<\/strong><\/h5>\n