eInfra requirements local systems

This tutorial describes the IT requirements to provision the PhenoMeNal e-infrastructure or parts thereof in your local compute center or IT-infrastructure. This is a living document that evolves over time. Provisioning PhenoMeNal VRE on a local system requires either a dedicated standalone server, or an OpenStack (www.openstack.com) private cloud installation.

E-infrastructure requirements for deploying PhenoMeNal Cloud Research Environments (CRE) to local systems

Hardware Requirements

The hardware requirements needed to run the PhenoMeNal e-infrastructure in your local compute environment obviously depend on the tasks that you intend to perform. A small test version of PhenoMeNal can run on just a single node whereas installations for large scale data processing will require state of the art compute clusters with tens, hundreds or thousands of nodes, depending on the scale. As soon as sufficient data on the scaling of PhenoMeNal compute workflows have been gathered, we will provide more detailed information in this location.

Contextualization

Provisioning of PhenoMeNal VREs consists of running cloud contextualization tools to spawn compute nodes, configure networks, attach storage, install the necessary software etc. In PhenoMeNal we use Terraform (http://terraform.io) to provision the host clouds, Ansible (http://www.ansible.com) to provision the instances and Packer (http://packer.io) to create the immutable images.

Requirements for OpenStack

OpenStack provides the API endpoints that Terrafrom needs to talk to, in order to fire up the infrastructure. A detailed list of these endpoints is available on: https://www.terraform.io/docs/providers/openstack/index.html. In order to install OpenStack, please refer to the extensive OpenStack documentation at www.openstack.com.

Requirements for standalone server

Information for how to provision PhenoMeNal VRE on a standalone server will be available soon.

References

• Rocca-Serra, P., Salek, R.M., Arita, M., Correa, E., Dayalan, S., Gonzalez-Beltran A., Ebbels T. et al. (2016) Data standards can boost metabolomics research, and if there is a will, there is a way. Metabolomics, 12(1):14. DOI: 10.1007/s11306-015-0879-3
• Merlet, B., Paulhe N., Vinson, F., Frainay, C., Chazalvie,l M., Poupin, N., Gloaguen, Y. et al. (2016) A Computational Solution to Automatically Map Metabolite Libraries in the Context of Genome Scale MetabolicNetworks. Front Mol Biosci., Feb 16; 3:2. DOI: 10.3389/fmolb.2016.00002
• Blaise, B.J., Correia, G., Tin, A., Young, J.H., Vergnaud, A.C., Lewis, M., Pearce, J.T. et al. (2016) Power Analysis and Sample Size Determination in Metabolic Phenotyping. Anal Chem. May 17; 88(10): 5179-88. DOI:10.1021/acs.analchem.6b00188
• Saccenti, E., Menichetti, G., Ghini, V., Remondini, D., Tenori, L., and Luchinat, C. (2016) Entropy–Based Network Representation of the Individual Metabolic Phenotype. J Proteome Res. Sep 2; 15(9): 3298-307. DOI:10.1021/acs.jproteome.6b00454
• Karaman, I., Ferreira, D.L., Boulange, C.L., Kaluarachchi, M.R., Herrington, D., Dona, A.C, Castagné, R. et al. (2016) A Workflow For Integrated Processing of Multi-Cohort Untargeted 1H NMR Metabolomics Data In Large ScaleMetabolic Epidemiology. J Proteome Res. 2016 Sep 15. DOI: 10.1021/acs.jproteome.6b00125

Contributors

Namrata Kale

Namrata S. Kale obtained her PhD in natural product chemistry from Indian Institute of Technology, New Delhi, India. She is currently a member of the Cheminformatics and Metabolism group at EMBL-EBI, where she works as a Project Manager for PhenoMeNal

Sijin He

Dr. Sijin He received a BSc in Mathematics (1st Class Hon) from University College London in 2007 and a MSc in Computing Science from Imperial College London in 2008. In August 2013, he received PhD in Computer Science at Imperial College London on cloud computing. He worked as a postdoctoral research in Data Science Institute. He is currently working as a full-stack software developer for PhenoMeNal H2020 project at EMBL-EBI.

Pablo Moreno

Ola Spjuth