HPCandBDSurvey2017V1

HPC and BD Survey Structure

The survey has several independent paths - i.e., you can hoose any combination of:

1. HPC hardware resources
2. BD layer 1: Repository (Figure 1 BDC white paper)
3. BD layer 2: Technological Infrastructure

[Note: layer 3, BD Applications, is not addressed in this survey, since this layer represents the outcome of using the resources in layers 1 and 2]

4. HPC Storage
5. Software environment
• HPC
• BD
6. Internet bandwidth
7. Consultation with Academic and Research Computing Systems (ARCS)

Terms used in the survey

• Explanation Description of the resource or service being considered
• Future link at HPC & BD wiki The documentation at HPC & BD Wiki is not yet available
• On-premise shared The resource is located at NJIT, and is shared amongst its users
• Off-premise shared The resource is not located at NJIT, and is shared amongst its users
• Off-premise dedicated The resource is not located at NJIT, and is dedicated to the researcher that purchased the resource
• External The resource is publically-available (e.g., at a national supercomputing center); successful proposal by researcher is required
• Storage
  • AFS, NFS distributed filesystems; scratch space [ future link at HPC & BD wiki ]
  • HDFS Hadoop Distributed File System

1. HPC computational resources

1.1 Assessment of adequacy/deficiency/improve of computational resources

1.1.3 Questions

• Rating of current adequacy / inadequacy
• If inadequate, how to improve
• Increase shared on-premise - resource use restrictions apply
• Increase dedicated on-premise - researcher purchases resources
• Increase shared off-premise - resource use restrictions apply
• Increase dedicated off-premise - researcher purchases resources
• Use external grid computing, e.g., Xsede

1.1.4 HPC Documentation

• Rating of adequacy
• 0-5 with 5 = excellent

2. HPC Storage

2.1 Assessment of adequacy/deficiency/improve

2.1.3 Questions

• Is your current allocation sufficient
• By what percentage should your current allocation be increased for that allocation to be sufficient
• 12 months from now
• 24 months from now
• If inadequate, how to improve
• researcher purchase on-premise resource
• researcher purchase off-premise resource
• other

2.1.4 Storage Documentation

• Rating of adequacy

3. BD layer 1: Repository

3.1.1 Resource: Shared storage for: raw data; metadata, markup data; analysis results; models, views, tables; forms, animations; workflow templates, provenance data

3.1.3 Questions

• Rating of current adequacy / inadequacy
• If inadequate, how to improve
• increase on-premise shared
• increase on-premise dedicated - researcher purchase
• increase off-premise shared
• increase off-premise dedicated - researcher purchase

4. BD layer 2: Technological Infrastructure

4.1 Assessment of adequacy/deficiency/improve of BD resources

4.1.3 Questions

• Rating of current adequacy / inadequacy
• If inadequate, how to improve
• increase on-premise shared
• increase on-premise dedicated - researcher purchase
• increase off-premise shared
• increase off-premise dedicated - researcher purchase

5. Software environment

5.1 Explanation:

The software environment is the combination of applications - opensource and commercial, libraries, utilities, and modules for setting the user's environment for specific software.

A listing of the available HPC software is at Software Modules.

5.2 Rating of overall software environment

• 0-5 with 5 = excellent
• In the following form, please list software not listed in the "Software Modules" in the link above that you would like to be available, and indicate a) whether free or not; and b) your expected use of it
• Comments on how the software environment could be improved

5.3 Software documentation

• 0-4 with 4 = excellent

6. Internet bandwidth

6.1 Explanation:

Internet bandwidth is the capacity of NJIT's connection to and from the Internet.

6.2 Rating of overall Internet bandwidth, including Internet 2

• 0-4 with 4 = excellent
• Comments on Internet bandwidth

7. Consultation

7.1 Explanation:

Consultation is communication with Academic and Research Computing Systems (ARCS) staff on topics such as getting started, problems encountered when running jobs, optimizing throughput, running parallel jobs, and managing disk space.

7.2 Rating of consultation

• 0-5 with 5 = excellent
• Comments on how the software environment could be improved