For a while I wanted to publish the research capabilities of Professor Jeremy Hajek for the Cloud and Operations research group in the Information Technology and Management Program Program at the School of Applied Technology.

Current Hardware Platform - as of 10/26/17

Current applications include:

  • 4 node Hadoop 2.6.5 Research Cluster
  • 5 node Hadoop 2.8.5 Research Cluster
    • 197 GB RAM and 47 processors available.
  • 1 Pfsense 2.4 VPN router for access to the cluster
  • 1 Instance of Cobbler 2.8.x for network based installs
  • Use of Riemann for metric and event routing
  • 1 instance of Grafana for visual inspection of metrics
  • 2 instances of Docker running on Ubuntu 16.04 LTS

Current Hardware

Software Support

  • Supported Software
    • MapReduce
    • Spark
    • SparkR
    • SparkQL


AppendixForITResroucesABET.pdf from 05/15

HP Eucalyptus 4.0.2 Cloud Computing Stack

Updated: 08/10/2016
Updated: 12/02/2016</p>

Eucalyptus is a private cloud platform now owned by HPE. Similar products would include OpenStack. Eucalyptus made the choice to make itself Amazon Web Servicescompatible. Though they lag behind in some cases, but the core functionality and syntax makes this still a useful product.

Our capacity is over 100 IPs supporting potentially over 200 virtual machine instances. There are currently 4 node controllers (using Linux KVM), 1 TB of object based storage (Object Storage Gateway) 500 GB of object based storage and 750 GB of Elastic Block based storage (EBS) 11 TB of Elastic Block based storage (EBS) using Western Digital Red drives. The HP Eucalyptus cloud supports similar function and control as AWS cloud. Capabilities include:

  • Elastic Load-balancing
  • Security Groups (Firewall)
  • RSA key based authentication (no passwords ever)
  • Cloud-watch (internal metrics)
  • Auto-scaling groups
  • University VPN access has been configured and used


There is a matrix of AWS API support as well that I configured and published:


There are two choices of instances that are currently available - Centos 6.8 and Ubuntu 14.04.5

The most recent course to run using this system was ITMT 430 Undergraduate Capstone Course which spent a semester building and deploying salable applications. Services are self-serve via a portal or via command line tool kit called Euca2ools. Current status requires a username password to be created for access - Active Directory integration is available and I would like to pursue this over the winter break of 2016 to have ready for Spring of 2017 and expand the availability to a larger group.

Location IIT Rice Campus

  • Incheon - Dell PowerEdge 2970 12 cores 32 GB of memory 1 TB of 15K SCSI local storage
  • Alexandria - Dell PowerEdge R710 8 cores 96 GB of memory 3 TB of local storage
  • Quantas - Dell PowerEdge R510 8 cores 32 GB of memory 3 TB of local storage
  • Nebelet - Dell PowerEdge R510 8 cores 32 GB of memory 3 TB of local storage
  • 1 Dell PowerEdge 1850
  • White Box Cloud Controller Intel G850 Processor 16 GB of memory 11 TB of storage
  • -> Cobbler server 1 Dell PowerEdge SC 1425


Deployment and Operations Research Cluster

Provides research capabilities into CEPH distributed file system clusters and MAAS (Metal as a Service) clustering for application install.

16 Dell Poweredge 1425 40 GB hard drive 4 GB of memory 2 Single core Xeon processors

Eucalyptus Private Test Cloud

Test Cloud 1 is used for the practice of installation and upgrade testing for production cloud, using Centos 6

  • 1 WhiteBox servers 8 cores 16 GB memory 1 TB of local storage
  • 3 ASUS RS100-X7 4 cores each ΓÇô 8 GB memory 1 TB local storage


Test Cloud 2 is used for research into pass-through via VFIO in the Linux Kernel for use in on demand GPU cloud computing resources.

  • 1 White Box servers 8 cores 16 GB memory 1 TB of local storage
  • 3 ASUS RS100-X7 4 cores each ΓÇô 8 GB memory 1 TB local storage
  • 1 GPU cluster for GPU research ΓÇô Containing for AMD Radeon
  • 3 white box storage systems.


Actual Image Located at the Wheaton Rice Campus - Room 242