1. From VPH-Share to PL-Grid: Atmosphere as an Advanced Frontend
to Cloud Resources in Research Infrastructures
Marian Bubak1,2, Bartosz Baliś1,2, Tomasz Bartyński1, Tomasz Gubała1, Daniel Harężlak1, Marek Kasztelnik1,
Maciej Malawski1,2, Jan Meizner1, Piotr Nowakowski1, Bartosz Wilk1, Paweł Suder1
1 ACC Cyfronet AGH, ul. Nawojki 11, Kraków, Poland
2 AGH, Department of Computer Science, al. Mickiewicza 30, Kraków, Poland
Atmosphere is a platform for developers and end users, supporting efficient development and execution of eScience applications on federated clouds.
4. The new service is registered and can be shared with other platform users
Service instance
OpenLab v1 (development)
Hardware: 4 cores,
8 GB RAM, 30 GB HDD
Compute Cloud
OpenLab v1
MySpine v22
GIMIAS
ViroLab DRS
OpenLab v1
The developer instructs the platform to package and upload a new service
Developer
The user can interact with the service. After the session, the instance can be shut down
Scientist
The developer selects and instantiates a service from an existing application or a “raw” OS image
The scientist browses the available services and requests access to a given service
Atmosphere creates a virtual machine and provides the required quantity of hardware resources.
Service instance
Ubuntu Linux (OS template)
Hardware: 4 cores,
8 GB RAM, 30 GB HDD
Service instance
OpenLab v1
Hardware: as required by the application
OpenLab v1
Authorized users:
The service’s owner approves the sharing request
The scientist requests the cloud platform to create an instance of the service
2’. The developer logs into the machine and installs software
Scientific objectives
Research on cloud computing model for complex scientific applications
Resource management for services on heterogeneous resources
Billing and accounting models for cloud computing
Procedural and technical aspects of efficient and secure data storage, transfer and processing
Component dependency management, composition and deployment
Applications
Virtual Patients for medical education
@neurIST – simulation of cerebral aneurysms
ARTreat – arthrosclerosis simulation
– dementia research
ViroLab – HIV therapy decision support system
VPHOP – modeling of osteroporosis
OpenLabirynth – medical training courses
MySpine – lumbar spine simulation
EUHeart – cardiovascular modeling
Atmosphere in numbers
Over 160 services registered
50–100 service instances run daily
More than 250 virtual machine templates
1000 requests for instance creation per month
100+ registered users from VPH community
VM lifetime from 15 minutes to over 60 days
30% of users with developer access
Features of Atmosphere
A layer of abstraction over cloud-based virtual machines
A way to port scientific applications to the cloud
Automatic selection of the best hardware resources to deploy applications
Automatic load balancing allows applications to scale up as needed
VM image synchronization for hybrid clouds
A billing model for commercial clouds
Integration with Taverna, HyperFlow, Pumpkin
Applications supported:
Linux-based SOAP/REST services
Web applications
rich GUI clients running under MS Windows
Integration with the VPH and PL-Grid ecosystem:
authentication, authorization, sharing
data management
Hybrid Cloud Platform
VPH-Share/PL-Grid Portal
VPH-Share Cloud Site at Cyfronet (OpenStack)
Cloud Management Portlets
VPH-Share Cloud Site at UNIVIE (OpenStack)
Secure RESTful API
(Cloud Facade)
Atmosphere Core Services Host
Amazon Elastic Compute Cloud (EC2) – EU Region
Atmosphere Core
PL-Grid Cloud Site at Cyfronet (OpenStack)
References
P. Nowakowski, T. Bartyński, M. Bubak, T. Gubała, D. Harężlak, M. Kasztelnik, M. Malawski, J. Meizner, Development, Execution and Sharing of VPH Applications in the Cloud with the Atmosphere Platform, VPH 2014 Conference, Trondheim, Norway 2014.
M. Bubak, M. Kasztelnik, M. Malawski, J. Meizner, P. Nowakowski, S. Varma: Evaluation of Cloud Providers for VPH Applications. CCGRID 2013:
M. Malawski, T. Gubała, M. Bubak: Component-based Approach for Programming and Running Scientific Applications on Grids and Clouds. The International Journal of High Performance Computing Applications, 26(3) : 2012