RNP Cloud Infrastructure model, services and challenges

Cloud Federation & Open
Science Cloud at cross-
regional level
RNP Cloud Infrastructure model,
services and challenges
José Luiz Ribeiro Filho
Cloudscape Brasil 2016
Porto Alegre – 7/7/2016

Summary
 Panel Objectives
 The European Open Science Cloud Model
 The European Open Science Cloud - Infrastructure and Services
 The Brazilian Cloud model
 The Brazilian Cloud model - Implementation and Service Definition
 RNP Cloud Services Status
 Elastic Compute
 Cloud Storage
 Web Conference
 Cloud E-mail
 Major Challenges
 Final Remarks

Panel Objectives
Federation of cloud resources is the way forward to address the free movement
of data for reasons other than the protection of personal data and by avoiding
vendor lock-in. The actual need is to have computing resources aside data stores
to reduce the latency and optimize the usage of the network infrastructure, and
to replicate and synchronize data to make services and applications more
resilient.
Cloud marketplaces are the business equivalent of a cloud federation by offering
customers the access to a wide variety of commercial services within the same
platform. Capitalizing and leveraging upon the Open Science Cloud within the
EU’s DSM strategy is a priority policy action to define a public, open cloud
infrastructure that all scientific researchers can use in an integrated way,
removing silos.
How the European Open Science model could be adopted in Brazil, in order to
establish a public cloud infrastructure for the Brazilian science community?

The European Open Science Cloud model
https://zenodo.org/record/16001/files/EUSciCloud-March2015.pdf
The model aims to "enable digital science by introducing IT as a Service to the
public research sector in Europe".
It is based on a "hybrid model that brings together public research organizations
and e‐Infrastructures with commercial suppliers to build a common platform
offering a range of services to Europe’s research communities".
Helix Nebula has demonstrated the potential of a hybrid model bringing
together service providers, research organizations, data providers and
publicly funded e‐infrastructures to support and transform publicly funded
research into data driven knowledge that is valuable to the wider research
community and downstream industries. The stakeholders have federated
their efforts and resources permitting a sub‐set oft he service providers to
develop a first product called HNX4 that is being marketed in a range of
business sectors.

The European Open Science Cloud model
The technical model offers "an extended but coherent set of services and tools,
organized into layers, which must be available to meet the specific needs of each
community and integrated via interfaces defined by open standards".
• should be able to act as the incubator for new businesses and scientific
activities
• rapidly expand the available capacity by making use of commercial service
providers offering commodity compute and data services as part of the
hybrid cloud model
• the overall management and control of the European Open Science Cloud
would be the task of the public research organizations so that a relationship
of trust can be established with the research communities while ensuring
non‐discrimination, transparency and equal treatment for all commercial
suppliers.

The European Open Science Cloud
infrastructure and services
The proposal has 3 distinct layers of services:
L1. European and international networks (GEANT); services for identity management and federation
across all European research and education institutions and integrated with other regions of the world
(eduGAIN);
L2. A group of facilities to provide cloud and data services of general and widespread usage. Linking
data services with cloud computing capacity to offer data analysis platforms will present users with a
comprehensive environment supporting the full lifecycle of science workflows.
L3. Software services and tools to provide value‐added capabilities to the research communities, in a
managed repository:
a. The tools to provide those research communities that have access to large sets of resources the
ability to federate and integrate those resources and to operate them for their community,
potentially sharing with other communities;
b. Tools to help build applications: e.g. tools to manage data, storage, workflows, visualization and
analysis libraries, etc.
c. Tools and services to allow researchers to integrate everyday activities with the European Open
Science Cloud: collaborative tools and services; office automation, negotiated licensing
agreements etc.;
d. Tools to help research communities engage the general public as citizen scientists.

The European Open Science Cloud
infrastructure and services
Example:
The following is the portfolio of services already provided by CERN to its users:
• A virtual multi‐tenant compute environment to provision and manage networks of virtual machines
on‐demand;
• A ‘dropbox’ style service for secure file sharing over the internet based on the ownCloud platform;
• A point‐to‐point reliable, automated file transfer service for bulk data transfers.
• The zenodo OpenAIRE compliant open access repository for publications and supporting data and
software allowing users to create and control their own digital libraries.
• A long‐term archiving service;
• Integrated Digital Conferencing tools allowing users to manage their conferences, workshops and
meetings (the digital conferencing tools already include commercial services such as Vidyo).
• Training services: Online training material will be made available.
The services are accessible in a single sign‐on (SSO) manner supported by a federated identity
management system and initial integration with eduGAIN has been performed.
A Service Portal, based on the servicenow commercial product, provides a common support
interface for the users.

8
European SDI Architecture Model and
Federated Infrastructure Components
http://www.uazone.org/demch/presentations/cts2014tutorial-cloud-federation.pdf

• The hybrid model allows RNP to act as a public (commercial) contract
broker for “low risk” services
• Through the community model, universities and research institutes can
offer their own data center infrastructure (in full or partially)
• The federated model ensures the required level of security and trustiness
among all the partners as well as to the users through RNP’s federated
services
• This strategy is also an answer to the geographical distribution requirement
for a redundant and reliable cloud
The Brazilian Cloud model

The Brazilian Cloud model
Implementation and Service Definition

11
Rede Ipe
Redecomep
eduroam
Compute@RNP
NAF@RNP
TBD
Cloudstack
Openstack
Fogbow
TBD
CAFe
.periodicos – CAPES
LiNEA
MConf
Brasilian SDI Architecture Model and
Federated Infrastructure Components

RNP Cloud Services Status – Elastic Compute
• The Elastic Compute service aims to provide virtual machines to the R&E
community, allowing researchers to accelerate tests and simulations, which
today depends of the acquirement of new hardware, which in some cases
impact their research.
• The service plans to start the pilot in the first half of 2016. RNP chooses Apache
CloudStack as the software platform that will provide the service. The project
team has ended the deployment of all systems and finished the test and
homologation phase.
• The service will support federated authentication by default and will provide a
variety of service offerings, like compute offerings since 1vCPU and 512MB of
RAM to 6vCPU and 12GB of RAM, disk offerings from 10 GB to 2 TB of capacity,
network offerings with firewall, load balance, DNS, NAT, VPN and others.

RNP Cloud Services Status – Cloud Storage
• CNC is a platform based on open source technologies and developed by a
Brazilian research group, funded by RNP, the pilot will stats in the first quarter
of 2016, for now the software is in final test phase.
• The service has a federated self-service portal and the users can upload,
download, visualize, share and synchronize their files. Additionally the service
has federated desktop and mobile apps and a federated administration portal
which administrators can manage quotas, users and groups in a federated
environments and visualize administrative reports and metrics.
• During the test and experimental phase many enhancements were
implemented, such like implement a multi-tenant and multi-container model
using swift as a backend for OwnCloud, free iOS client, etc.

RNP Cloud Services Status - Web Conference
• Mconf is a platform based on open source technologies and developed by a
Brazilian research group, funded by RNP, the platform is running as a pilot for
RNP’s staff since October 2014 and will replace the old platform in a near
future.
• The service has a self-service portal and the users can create communities,
manage other users on communities, delegate administration of communities,
record and manage web conference sessions and more. Additionally all
activities are logged, bringing accounting capabilities.
• During the pilot and experimental service many enhancements were
implemented, such like a better user interface, a Web RTC module to improve
the audio quality, a mobile app, etc.

RNP Cloud Services Status – Cloud E-mail
• The Cloud e-mail service is under construction and the main goal of the service is
to provide a cloud based e-mail service that will allow the R&E community to host
e-mail boxes to their staff, lecturers, researchers and post graduate students.
• The service must to have a self-service model where each institution that
subscribe the service will be able to set their own domains, manage their users,
connect your own LDAP database, manage spam, white and black lists, customize
their web interface for the users, manage message features like attachment size,
etc.
• Currently RNP has defined the software
platform which will provide the service.
The project team has ended the
deployment of all systems and finished
the test and homologation phase. The
next steps are: finish the business and
management models and start to migrate
the accounts from pilot R&D institutions
to validate the processes and procedures
and improve the service.

Major Challenges
General
• Legal frameworks in Europe and Brazil
• Data protection (exchange and security)
• Intercontinental networking
• Specific (in Brazil)
• Community/shared model implementation
• Accounting model
• Funding and sustainability model
• Governance and management model

Major Challenges
• Stakeholders
• Public R&E institutions can be, at the same time or separately,
customers and providers of infrastructure and services;
• Research groups, that contribute with their scientific
application’s requirements;
• R&E funding agencies, that regulates the current research-
funding paradigm, which need to be changed based in the
cloud paradigm;
• RNP that has the coordination role, being responsible for the
cloud architecture definition, contracting and operating its
services.
• Users
• Institutional users (IT managers of R&E institutions) – Use
traditional workloads and architectures;
• Individuals (e.g. researchers and their labs) – Use dynamic and
non-standard workloads and architectures.

Final Remarks
• The hybrid, community and federated strategy was chosen as the
most flexible and suitable for the Brazilian NREN operation and
funding models;
• The cloud implementation and services selected to be initially
offered are being deployed, but aren’t yet in production;
• RNP believes that the Brazilian academic cloud will increase the
security and the national sovereignty. It will also reduce current
fragilities that many public R&E institutions demonstrate regarding
the safety of an ever-increasing amount of strategic and vital
information, as a consequence of a lack of adequate e-
infrastructure services and support;
• Both the European and the Brazilian Cloud models are compatible
and allow integration of services and collaboration of both science
communities.

José Luiz Ribeiro Filho
Diretor de Serviços e Soluções – RNP
jose.luiz@rnp.br

RNP Cloud Infrastructure model, services and challenges

Related slideshows

More Related Content

RNP Cloud Infrastructure model, services and challenges