XVII
Workshop on Clouds and Applications

Migração de Sistemas Monolíticos para Microserviços: Uma Abordagem Híbrida sem Downtime

Augusto Flávio Arraes (Instituto Federal de Educação, Ciência e Tecnologia do Ceará), Américo Sampaio (UNIFOR), Otavio Medeiros (UNIFOR), Nabor Mendonça (UNIFOR)

Estratégias para Implementação de Coreografia de Microsserviços

João Pedro Fernandes (UTFPR), Edson Tavares de Camargo (UTFPR)

Distributed Cloud: A Modern Paradigm Towards 5G Networks

Moacyr Martucci Junior (USP)

Abstract: The move from 4G to 5G (Fifth Generation of Mobile Networks) promises a very significant jump in terms of network performance, with the support of millimeter waves, small cells, massive Multiple-Input and Multiple-Output, distributed cloud computing etc. being deployed to lower latency and increased throughput. The combination of these technologies is expected to deliver a 10X decrease in latency – as low as 1 ms – and a 10X increase in speed – higher than 1 Gbps. Network slicing, a key aspect of 5G networks, will be important for the edge. Internet Service Providers (ISPs) can implement slicing without 5G, but it will likely become much more prevalent with 5G and its emerging specifications that mandate the partitioning of the data, control and management planes to separate environments that can be created. Thus, serving individual customers or deliver specific services, giving ISPs the opportunity to more easily support multi-tenancy, individual customers and use cases in order to meet each slice´s unique Service-Level Agreements. The advent of 5G is helping to drive demand for distributed edge architectures because of the technology advancements that it promises. The Internet Protocol (IP) network will need to support data, control, and management plane separation across a multi-site network fabric as a result. All these changes are leading to new requirements for the IP network and network fabric. For the IP network, this means that traffic will need to be processed at the edge in thousands of mini -and micro- data centers, some of which may be located in the Radio Access Network. The IP network, ideally leveraging the flexibility and cost-effectiveness of white boxes, will need to support a network fabric that can make this large number of sites appear, logically, as a more manageable set of network fabrics. The white boxes will need to offer some sort of SDN (Software-Defined Networking) control and ideally some sort of network fabric that will make a large number of sites appear, logically, as a set of fewer network fabrics. Using Application Programming Interfaces, these fabrics will need fabric-wide programmability, helping ISPs improve their operational efficiency at scale. SDN brings easier network automation and service orchestration to improve operational efficiency. 5G Network slicing will drive the need for the IP network to support data, control, and management plane separation across a multi-site network fabric, which should have a central view of network state, but with that knowledge and intelligence distributed across the whole fabric and then throughout the distributed cloud.

Bio: Full professor at the Department of Engineering and Digital Systems of Escola Politécnica of Universidade de São Paulo and the Coordinator of the Institute for Studies Brazil Europe, a Research Support Nucleus under the Research Pro-rectory of the USP. Martucci is also the Brazilian National Contact Point in Program Horizon 2020 and the NCP in the program ideas to ERC (European Research Council) to European Commission and participates in the Brazil-EU group within the Action "5G and the Future of Mobile telecommunications in Brazil and Europe" at the Ministry of Communications and Ministry of Planning in Dialogue Sector BR-EU. His main interest areas are Information Systems, Open Systems, Computer Networks and Automation, 5G and Future Internet and Internet of Things with applications in smart cities.

RW-Through: A Data Replication Protocol Suitable for GeoDistributed and Read-Intensive Workloads

Wagner Barreto (UTFPR), Ana Cristina B. K. Vendramin (UTFPR), Mauro Fonseca (UTFPR)

Impacto de LoRaWAN no Desempenho de Plataformas de IoT baseadas em Nuvem e Névoa Computacional

Ivan Zyrianoff (UFABC), Alexandre Heideker (UFABC), Dener Silva (UFABC), João Henrique Kleinschmidt (UFABC), Carlos Kamienski (UFABC)

Heap Allocator: Uma Polıtica de Alocação de Máquinas Virtuais em Ambiente de Computação em Nuvens baseada em Heap com Prioridade

Matheus Magalhães de Carvalho (Universisdade Estadual do Ceará - UECE), Francisco Vieira (UECE), Filipe de Matos (Universidade Federal do Ceará), Rafael Lopes Gomes (UECE), Joaquim Celestino (UECE)