HCC's Crane Supercomputer Vs. Amazon EC2

CraneVsEc2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Amazon EC2 Instance\n",
    "We will assume an Amazon EC2 instance of type `m4.xlarge`.  It has these characteristics:\n",
    "* 4 cores\n",
    "* 16 GB of RAM (4GB / core)\n",
    "* $0.239 per Hour\n",
    "\n",
    "Crane also has 4 GB of Ram per core."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate Computation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Crane has 452 nodes at 16 Cores per node\n",
    "crane_cores = 452*16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Calculate how many instances are needed to simulate Crane\n",
    "num_instances = crane_cores / 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hourly: $432.11\n",
      "Daily: $10,370.69\n",
      "Monthly: $311,120.64\n",
      "Annual: $3,785,301.12\n"
     ]
    }
   ],
   "source": [
    "# How much does it cost per hour for Crane to be on EC2, $0.239 per hour\n",
    "hourly_cost = num_instances * 0.239 \n",
    "print 'Hourly: ${:0,.2f}'.format(hourly_cost)\n",
    "print 'Daily: ${:0,.2f}'.format(hourly_cost*24)\n",
    "print 'Monthly: ${:0,.2f}'.format(hourly_cost*24*30)\n",
    "print 'Annual: ${:0,.2f}'.format(hourly_cost*24*365)\n",
    "compute_annual_cost = hourly_cost*24*365"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate Storage\n",
    "Amazon S3 has tiered pricing.  But for simplicity of the calculations, we will only consider the top, and least expensive tier of $0.0275 per GB / Month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Crane Storage in GBs: 1,486,848\n"
     ]
    }
   ],
   "source": [
    "# Crane has 1452 TB of Lustre\n",
    "crane_gb = 1452 * 1024\n",
    "print 'Crane Storage in GBs: {:0,.0f}'.format(crane_gb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Monthly: $40,888.32\n",
      "Annual: $490,659.84\n"
     ]
    }
   ],
   "source": [
    "storage_price_per_month = crane_gb * 0.0275\n",
    "print 'Monthly: ${:0,.2f}'.format(storage_price_per_month)\n",
    "print 'Annual: ${:0,.2f}'.format(storage_price_per_month*12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total Annual: $4,275,960.96\n"
     ]
    }
   ],
   "source": [
    "total_annual = (storage_price_per_month * 12) + compute_annual_cost\n",
    "print 'Total Annual: ${:0,.2f}'.format(total_annual)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}

Networking costs intentionally neglected. Operating includes FTE, power/cooling, space, racks and infrastructure.

Of course, one would need perhaps even more networking, and FTE in different role, with AWS as well.