parsing-experiment-transient.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "language_info": {
      "name": "python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "version": "3.6.7-final"
    },
    "orig_nbformat": 2,
    "file_extension": ".py",
    "mimetype": "text/x-python",
    "name": "python",
    "npconvert_exporter": "python",
    "pygments_lexer": "ipython3",
    "version": 3,
    "kernelspec": {
      "name": "python36764bit06bf6d74bc974779a98a6ef0a50c382d",
      "display_name": "Python 3.6.7 64-bit"
    },
    "colab": {
      "name": "parsing-experiment-transient.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RGmatESPar_Z"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/pacslab/serverless-temporal-perf-modeling/blob/main/parsing-experiment-transient.ipynb\" target=\"_parent\">\n",
        "  <img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/>\n",
        "</a>\n",
        "\n",
        "# Parsing Multiple Files for Modeling Transient Behaviour of Serverless Computing\n",
        "\n",
        "In this file, we aim to average out several executions of the same workload pattern to get the average behavior of the platform.\n",
        "\n",
        "If you are running this notebook from environments like **Google Colab**, uncomment the following cell in your first execution to download everything from the [Github repository](https://github.com/pacslab/serverless-temporal-perf-modeling)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o_uu_GY5a09j"
      },
      "source": [
        "# If running from Google Colab, uncomment these to fetch the data from Github:\n",
        "# !git init\n",
        "# !git remote add origin https://github.com/pacslab/serverless-temporal-perf-modeling.git\n",
        "# !git pull origin main\n",
        "# !pip install -r requirements.txt"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "tags": [],
        "id": "ZXlFOld2ar_a",
        "outputId": "5a248890-e647-40f1-b3c8-0e684794a495",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "%load_ext autoreload\n",
        "%autoreload 2\n",
        "# Imports\n",
        "\n",
        "import os\n",
        "\n",
        "import pacsltk.pacs_util as pacs_util\n",
        "import trans_exp_parser as exp_parser\n",
        "import trans_perf_model\n",
        "\n",
        "import pandas as pd\n",
        "\n",
        "pacs_util.prepare_matplotlib_cycler()\n",
        "\n",
        "# To avoid type 3 fonts: http://phyletica.org/matplotlib-fonts/\n",
        "from matplotlib import rcParams\n",
        "rcParams['pdf.fonttype'] = 42\n",
        "rcParams['ps.fonttype'] = 42\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from pandas.plotting import register_matplotlib_converters\n",
        "register_matplotlib_converters()\n",
        "\n",
        "import pacsltk\n",
        "print(pacsltk.__version__)"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.2.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5gpJQRPSar_f"
      },
      "source": [
        "csv_file_folder = os.path.join('results', 'exp1')\n",
        "csv_file_list = [i for i in os.listdir(csv_file_folder) if '.csv' in i]\n",
        "\n",
        "idle_mins_before_kill = 10\n",
        "step_seconds = 10\n",
        "\n",
        "tdfs = []\n",
        "dfcs = []\n",
        "for csv_filename in csv_file_list:\n",
        "    csv_file = os.path.join(csv_file_folder, csv_filename)\n",
        "\n",
        "    df = exp_parser.read_csv_file(csv_file)\n",
        "    df = exp_parser.preprocess_df(df)\n",
        "    tdf, time_idx = exp_parser.time_preprocess_df(df)\n",
        "\n",
        "    parse_results = exp_parser.parse_instance_info(df, idle_mins_before_kill)\n",
        "    # load results into local variables\n",
        "    locals().update(parse_results)\n",
        "\n",
        "    # tdf['instance_count'] = \n",
        "    df_counts = exp_parser.parse_counting_info(df, df_inst, step_seconds=step_seconds)\n",
        "    df_counts['rps'] = tdf['is_cold'].resample('{}S'.format(step_seconds)).count() / step_seconds\n",
        "    df_counts['p_cold'] = tdf['is_cold'].resample('{}S'.format(step_seconds)).mean()\n",
        "    df_counts['rt'] = tdf['client_elapsed_time'].resample('{}S'.format(step_seconds)).mean() / 1000\n",
        "\n",
        "    df_counts_smooth = df_counts.resample('T').mean()\n",
        "\n",
        "    # Append the results to be used later\n",
        "    tdfs.append(tdf)\n",
        "    dfcs.append(df_counts_smooth)"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ovyps7e6ar_i",
        "outputId": "9b40070e-f3f7-4397-e965-d9aba6d3bd15",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        }
      },
      "source": [
        "tdfs[0].head()"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>is_cold</th>\n",
              "      <th>inst_id</th>\n",
              "      <th>vm_id</th>\n",
              "      <th>elapsed_time</th>\n",
              "      <th>aws_duration</th>\n",
              "      <th>client_elapsed_time</th>\n",
              "      <th>client_start_time_dt</th>\n",
              "      <th>client_end_time_dt</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>client_start_time_dt</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0 days 00:00:00</th>\n",
              "      <td>True</td>\n",
              "      <td>sandbox-cba6dc</td>\n",
              "      <td>sandbox-root-CxJkRb</td>\n",
              "      <td>2032.337402</td>\n",
              "      <td>2033.02</td>\n",
              "      <td>2427.139521</td>\n",
              "      <td>2020-03-29 23:34:41.874080</td>\n",
              "      <td>2020-03-29 23:34:44.301220</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:00:03.056426</th>\n",
              "      <td>False</td>\n",
              "      <td>sandbox-cba6dc</td>\n",
              "      <td>sandbox-root-CxJkRb</td>\n",
              "      <td>1842.882568</td>\n",
              "      <td>1843.52</td>\n",
              "      <td>1931.809187</td>\n",
              "      <td>2020-03-29 23:34:44.930506</td>\n",
              "      <td>2020-03-29 23:34:46.862315</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:00:03.381933</th>\n",
              "      <td>True</td>\n",
              "      <td>sandbox-0accc0</td>\n",
              "      <td>sandbox-root-tH885o</td>\n",
              "      <td>1965.697266</td>\n",
              "      <td>1966.37</td>\n",
              "      <td>2528.977871</td>\n",
              "      <td>2020-03-29 23:34:45.256013</td>\n",
              "      <td>2020-03-29 23:34:47.784991</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:00:04.001833</th>\n",
              "      <td>True</td>\n",
              "      <td>sandbox-fb1c5c</td>\n",
              "      <td>sandbox-root-EO5dtO</td>\n",
              "      <td>2000.459961</td>\n",
              "      <td>2001.12</td>\n",
              "      <td>2599.175215</td>\n",
              "      <td>2020-03-29 23:34:45.875913</td>\n",
              "      <td>2020-03-29 23:34:48.475089</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:00:06.978754</th>\n",
              "      <td>False</td>\n",
              "      <td>sandbox-fb1c5c</td>\n",
              "      <td>sandbox-root-EO5dtO</td>\n",
              "      <td>1850.534912</td>\n",
              "      <td>1851.19</td>\n",
              "      <td>1943.536282</td>\n",
              "      <td>2020-03-29 23:34:48.852834</td>\n",
              "      <td>2020-03-29 23:34:50.796371</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        is_cold         inst_id                vm_id  \\\n",
              "client_start_time_dt                                                   \n",
              "0 days 00:00:00            True  sandbox-cba6dc  sandbox-root-CxJkRb   \n",
              "0 days 00:00:03.056426    False  sandbox-cba6dc  sandbox-root-CxJkRb   \n",
              "0 days 00:00:03.381933     True  sandbox-0accc0  sandbox-root-tH885o   \n",
              "0 days 00:00:04.001833     True  sandbox-fb1c5c  sandbox-root-EO5dtO   \n",
              "0 days 00:00:06.978754    False  sandbox-fb1c5c  sandbox-root-EO5dtO   \n",
              "\n",
              "                        elapsed_time  aws_duration  client_elapsed_time  \\\n",
              "client_start_time_dt                                                      \n",
              "0 days 00:00:00          2032.337402       2033.02          2427.139521   \n",
              "0 days 00:00:03.056426   1842.882568       1843.52          1931.809187   \n",
              "0 days 00:00:03.381933   1965.697266       1966.37          2528.977871   \n",
              "0 days 00:00:04.001833   2000.459961       2001.12          2599.175215   \n",
              "0 days 00:00:06.978754   1850.534912       1851.19          1943.536282   \n",
              "\n",
              "                             client_start_time_dt         client_end_time_dt  \n",
              "client_start_time_dt                                                          \n",
              "0 days 00:00:00        2020-03-29 23:34:41.874080 2020-03-29 23:34:44.301220  \n",
              "0 days 00:00:03.056426 2020-03-29 23:34:44.930506 2020-03-29 23:34:46.862315  \n",
              "0 days 00:00:03.381933 2020-03-29 23:34:45.256013 2020-03-29 23:34:47.784991  \n",
              "0 days 00:00:04.001833 2020-03-29 23:34:45.875913 2020-03-29 23:34:48.475089  \n",
              "0 days 00:00:06.978754 2020-03-29 23:34:48.852834 2020-03-29 23:34:50.796371  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JA8yMk-6ar_l",
        "outputId": "d6a28867-b14a-4b99-ac34-f9717e729595",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 203
        }
      },
      "source": [
        "dfcs[0].tail()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>instance_count</th>\n",
              "      <th>running_count</th>\n",
              "      <th>running_warm_count</th>\n",
              "      <th>idle_count</th>\n",
              "      <th>utilization</th>\n",
              "      <th>rps</th>\n",
              "      <th>p_cold</th>\n",
              "      <th>rt</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0 days 00:55:00</th>\n",
              "      <td>51.0</td>\n",
              "      <td>17.500000</td>\n",
              "      <td>17.500000</td>\n",
              "      <td>33.500000</td>\n",
              "      <td>0.343137</td>\n",
              "      <td>7.616667</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.059014</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:56:00</th>\n",
              "      <td>51.0</td>\n",
              "      <td>14.166667</td>\n",
              "      <td>14.166667</td>\n",
              "      <td>36.833333</td>\n",
              "      <td>0.277778</td>\n",
              "      <td>5.650000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.056468</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:57:00</th>\n",
              "      <td>51.0</td>\n",
              "      <td>6.833333</td>\n",
              "      <td>6.833333</td>\n",
              "      <td>44.166667</td>\n",
              "      <td>0.133987</td>\n",
              "      <td>3.066667</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.053080</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:58:00</th>\n",
              "      <td>51.0</td>\n",
              "      <td>1.666667</td>\n",
              "      <td>1.666667</td>\n",
              "      <td>49.333333</td>\n",
              "      <td>0.032680</td>\n",
              "      <td>1.033333</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.048078</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0 days 00:59:00</th>\n",
              "      <td>51.0</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>50.000000</td>\n",
              "      <td>0.019608</td>\n",
              "      <td>0.600000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.061483</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                 instance_count  running_count  running_warm_count  \\\n",
              "0 days 00:55:00            51.0      17.500000           17.500000   \n",
              "0 days 00:56:00            51.0      14.166667           14.166667   \n",
              "0 days 00:57:00            51.0       6.833333            6.833333   \n",
              "0 days 00:58:00            51.0       1.666667            1.666667   \n",
              "0 days 00:59:00            51.0       1.000000            1.000000   \n",
              "\n",
              "                 idle_count  utilization       rps  p_cold        rt  \n",
              "0 days 00:55:00   33.500000     0.343137  7.616667     0.0  2.059014  \n",
              "0 days 00:56:00   36.833333     0.277778  5.650000     0.0  2.056468  \n",
              "0 days 00:57:00   44.166667     0.133987  3.066667     0.0  2.053080  \n",
              "0 days 00:58:00   49.333333     0.032680  1.033333     0.0  2.048078  \n",
              "0 days 00:59:00   50.000000     0.019608  0.600000     0.0  2.061483  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kFQLQn_-ar_o"
      },
      "source": [
        "# Getting the average behaviour\n",
        "dfc = pd.DataFrame()\n",
        "for col_name in dfcs[0].columns:\n",
        "    tmpdf = pd.DataFrame()\n",
        "    for i in range(len(dfcs)):\n",
        "        tmpdf[col_name + \"_\" + str(i)] = dfcs[i][col_name]\n",
        "\n",
        "    dfc[col_name] = tmpdf.mean(axis=1)\n",
        "\n",
        "dfc.head()\n",
        "csv_filename = \"exp01-average.csv\"\n",
        "csv_file = os.path.join('results', csv_filename)\n",
        "dfc.to_csv(csv_file)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EcCc9m_9ar_t"
      },
      "source": [
        "mdf = dfc\n",
        "\n",
        "# How many minutes to predict into future\n",
        "pred_mins = 5\n",
        "\n",
        "warm_service_time = df[df['is_cold'] == False]['client_elapsed_time'].mean() / 1000\n",
        "cold_service_time = df[df['is_cold'] == True]['client_elapsed_time'].mean() / 1000\n",
        "idle_time_before_kill = 10*60\n",
        "max_t = 60\n",
        "\n",
        "trans_predictions = []\n",
        "for idx in range(mdf.shape[0] - pred_mins):\n",
        "    pred_rpss = mdf['rps'].iloc[idx:idx+pred_mins].values\n",
        "    init_server_count = int(mdf['instance_count'].iloc[idx])\n",
        "    props = None\n",
        "    for i in range(pred_mins):\n",
        "        arrival_rate = pred_rpss[i]\n",
        "        props, _ = trans_perf_model.get_sls_props_transient(arrival_rate, warm_service_time, \n",
        "                cold_service_time, idle_time_before_kill, max_t=max_t, \n",
        "                init_server_count=init_server_count)\n",
        "        init_server_count = int(props['avg_server_count'])\n",
        "        props['idx'] = mdf.index[idx]\n",
        "        \n",
        "    props['last_pred'] = pred_rpss[-1]\n",
        "    trans_predictions.append(props)\n",
        "\n",
        "pred_df = pd.DataFrame(trans_predictions)\n",
        "pred_df.set_index('idx', inplace=True)\n",
        "# shift to see predictions\n",
        "from datetime import timedelta\n",
        "pred_df.index = pred_df.index + timedelta(minutes=5)"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fINLgSo7ar_w"
      },
      "source": [
        "## Making Plots to Compare Model and Experiments"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GaqSNqx8ar_x"
      },
      "source": [
        "import matplotlib.ticker as mticker\n",
        "import datetime\n",
        "import math\n",
        "\n",
        "# Prepare for plots\n",
        "figs_folder = \"figs\"\n",
        "\n",
        "def get_fig_path(x): return (os.path.join(figs_folder, \"exp\" +\n",
        "                                          x + \".png\"), os.path.join(figs_folder, \"exp\" + x + \".pdf\"))\n",
        "\n",
        "def tmp_fig_save(fig_name):\n",
        "    paths = get_fig_path(fig_name)\n",
        "    plt.savefig(paths[0], dpi=300)\n",
        "    plt.savefig(paths[1])\n",
        "\n",
        "def plot_configs(ylab, is_texp=True):\n",
        "    plt.legend()\n",
        "    plt.tight_layout()\n",
        "    plt.grid(True)\n",
        "    plt.xlabel(\"Experiment Time (MM:SS)\")\n",
        "    plt.ylabel(ylab)\n",
        "    plt.gcf().subplots_adjust(left=0.17, bottom=0.22)\n",
        "    if is_texp:\n",
        "        pacs_util.fix_timedelta(1e-3)\n",
        "\n",
        "    def timeTicks(x, pos):\n",
        "        # d = datetime.timedelta(seconds=x/1e9)\n",
        "        # s = str(d)\n",
        "        # return \":\".join(s.split(\":\")[1:3])\n",
        "        remaining_seconds = x / 1e9\n",
        "        minutes = math.floor(remaining_seconds / 60)\n",
        "        seconds = math.floor(remaining_seconds - minutes * 60)\n",
        "        return f\"{minutes:02d}:{seconds:02d}\"\n",
        "\n",
        "    plt.gca().xaxis.set_major_locator(mticker.MultipleLocator(base=10*60*1e9))\n",
        "    plt.gca().xaxis.set_minor_locator(mticker.MultipleLocator(base=1*60*1e9))\n",
        "    formatter = mticker.FuncFormatter(timeTicks)\n",
        "    plt.gca().xaxis.set_major_formatter(formatter)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "tags": [],
        "id": "WQsG8p5Bar_0",
        "outputId": "c5325dea-1e69-4cdd-80e8-4baf619d25ab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 821
        }
      },
      "source": [
        "# Remove the parts we don't have a prediction for\n",
        "def make_plots(mdf, pred_df, prefix=None):\n",
        "    mdf = mdf[mdf.index >= pred_df.index.min()]\n",
        "\n",
        "    plt.figure(figsize=(4,1.8))\n",
        "    plt.plot(mdf['instance_count'], label='Experiment')\n",
        "    plt.plot(pred_df['avg_server_count'], label='Model Prediction')\n",
        "    plot_configs(\"Instance Count\")\n",
        "    if prefix is not None:\n",
        "        tmp_fig_save(prefix+\"_instance_count\")\n",
        "\n",
        "    plt.figure(figsize=(4,1.8))\n",
        "    plt.plot(mdf['rps'], label='Experiment')\n",
        "    plt.plot(pred_df['last_pred'].shift(-1), label='Model Prediction')\n",
        "    plot_configs(\"Requests / Sec\")\n",
        "    if prefix is not None:\n",
        "        tmp_fig_save(prefix+\"_rps\")\n",
        "\n",
        "    plt.figure(figsize=(4,1.8))\n",
        "    plt.plot(mdf['p_cold'], label='Experiment')\n",
        "    plt.plot(pred_df['cold_prob'], label='Model Prediction')\n",
        "    plot_configs(\"Cold Start Prob.\")\n",
        "    if prefix is not None:\n",
        "        tmp_fig_save(prefix+\"_p_cold\")\n",
        "\n",
        "    plt.figure(figsize=(4,1.8))\n",
        "    plt.plot(mdf['utilization'], label='Experiment')\n",
        "    plt.plot(pred_df['avg_utilization'], label='Model Prediction')\n",
        "    plot_configs(\"Utilization\")\n",
        "    if prefix is not None:\n",
        "        tmp_fig_save(prefix+\"_utilization\")\n",
        "\n",
        "    plt.figure(figsize=(4,1.8))\n",
        "    plt.plot(mdf['running_count'], label='Experiment')\n",
        "    plt.plot(pred_df['avg_running_count'], label='Model Prediction')\n",
        "    plot_configs(\"Num of Running\")\n",
        "    if prefix is not None:\n",
        "        tmp_fig_save(prefix+\"_running_count\")\n",
        "\n",
        "    plt.figure(figsize=(4,1.8))\n",
        "    plt.plot(mdf['idle_count'], label='Experiment')\n",
        "    plt.plot(pred_df['avg_idle_count'], label='Model Prediction')\n",
        "    plot_configs(\"Num of Idle\")\n",
        "    if prefix is not None:\n",
        "        tmp_fig_save(prefix+\"_idle_count\")\n",
        "\n",
        "make_plots(mdf, pred_df, prefix='01')"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x129.6 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x129.6 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x129.6 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x129.6 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x129.6 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x129.6 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UgRUlTjnjarg"
      },
      "source": [
        "# Download Figures\n",
        "\n",
        "This section will execute the necessary code for downloading the figures generated in this notebook."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OiWajDYSjZmX",
        "outputId": "870728ea-12d2-475c-d152-e9763355ad16",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 223
        }
      },
      "source": [
        "!zip figs.zip figs/*\n",
        "\n",
        "from google.colab import files\n",
        "\n",
        "files.download('figs.zip')"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "updating: figs/exp01_idle_count.pdf (deflated 2%)\n",
            "updating: figs/exp01_idle_count.png (deflated 6%)\n",
            "updating: figs/exp01_instance_count.pdf (deflated 2%)\n",
            "updating: figs/exp01_instance_count.png (deflated 6%)\n",
            "updating: figs/exp01_p_cold.pdf (deflated 2%)\n",
            "updating: figs/exp01_p_cold.png (deflated 6%)\n",
            "updating: figs/exp01_rps.pdf (deflated 2%)\n",
            "updating: figs/exp01_rps.png (deflated 6%)\n",
            "updating: figs/exp01_running_count.pdf (deflated 2%)\n",
            "updating: figs/exp01_running_count.png (deflated 5%)\n",
            "updating: figs/exp01_utilization.pdf (deflated 2%)\n",
            "updating: figs/exp01_utilization.png (deflated 6%)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_ae350bc4-a063-4bf7-ab5a-511364761ba0\", \"figs.zip\", 1711568)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}