gzip compression of resources is not enabled.

[AlexBravo]

Initial loading of Robonaut demo take > 30 sec and mostly bandwidth bound.
From Audit tab:
Enable gzip compression (33)
Compressing the following resources with gzip could reduce their transfer size by about two thirds (~7.4 MB):
three.min.js could save ~285 KB
ColladaLoader2.min.js could save ~52.1 KB
eventemitter2.min.js could save ~4.7 KB
roslib.min.js could save ~10.1 KB
ros3d.min.js could save ~30.9 KB
resources.robotwebtools.org/r2_description/meshes/Stanchion_Baseplate.dae could save ~45.4 KB
resources.robotwebtools.org/r2_description/meshes/Stanchion.dae could save ~518 KB
resources.robotwebtools.org/r2_description/meshes/Body.dae could save ~2.1 MB
resources.robotwebtools.org/r2_description/meshes/Body_Cover.dae could save ~144 KB
resources.robotwebtools.org/r2_description/meshes/Backpack.dae could save ~1.5 MB
resources.robotwebtools.org/r2_description/meshes/Neck_Lower.dae could save ~150 KB
resources.robotwebtools.org/r2_description/meshes/Neck_Upper.dae could save ~152 KB
resources.robotwebtools.org/r2_description/meshes/Head.dae could save ~604 KB
resources.robotwebtools.org/r2_description/meshes/Left_Shoulder_Upper.dae could save ~119 KB
resources.robotwebtools.org/r2_description/meshes/Left_Shoulder_Lower.dae could save ~53.8 KB
resources.robotwebtools.org/r2_description/meshes/Left_Upper_Arm.dae could save ~156 KB
resources.robotwebtools.org/r2_description/meshes/Left_Elbow.dae could save ~73.3 KB
resources.robotwebtools.org/r2_description/meshes/Left_Forearm.dae could save ~188 KB
resources.robotwebtools.org/r2_description/meshes/Left_Palm.dae could save ~94.6 KB
resources.robotwebtools.org/r2_description/meshes/Left_Thumb_Carp.dae could save ~55.5 KB
resources.robotwebtools.org/r2_description/meshes/Left_Thumb_MtCar.dae could save ~77.4 KB
resources.robotwebtools.org/r2_description/meshes/Thumb_Proximal.dae could save ~66.4 KB
resources.robotwebtools.org/r2_description/meshes/Thumb_Dist.dae could save ~96.4 KB
resources.robotwebtools.org/r2_description/meshes/Finger_Proximal.dae could save ~44.1 KB
resources.robotwebtools.org/r2_description/meshes/Finger_Mid.dae could save ~41.2 KB
resources.robotwebtools.org/r2_description/meshes/Finger_Dist.dae could save ~79.3 KB
resources.robotwebtools.org/r2_description/meshes/Right_Shoulder_Upper.dae could save ~111 KB
resources.robotwebtools.org/r2_description/meshes/Right_Shoulder_Lower.dae could save ~50.8 KB
resources.robotwebtools.org/r2_description/meshes/Right_Upper_Arm.dae could save ~152 KB
resources.robotwebtools.org/r2_description/meshes/Right_Elbow.dae could save ~53.6 KB
resources.robotwebtools.org/r2_description/meshes/Right_Forearm.dae could save ~188 KB
resources.robotwebtools.org/r2_description/meshes/Right_Palm.dae could save ~103 KB
resources.robotwebtools.org/r2_description/meshes/Right_Thumb_MtCar.dae could save ~91.3 KB

[rctoris]

Good call! I looked into it and it looks like AWS supports serving GZipped files on their CDN platform (cloud front) which is what I am using to server the files. I'll put this on my stack to try it out and see if it works.

[jihoonl]

Is this still valid issue?

[rctoris]

Yes and no. The answer is yes it does support it, but I have not updated the PR2 meshes to do it. We have done and tested it with our CARL robot and local collada models we use. We had to make a script to do it since it's a pain to do manually. Here is ours, we basically would need one for the PR2 but I can't test/verify since we no longer have one in our lab :(

!/bin/bash

TMP="/tmp/aws-rail-collada-models-sync"
SYNC="sync"
RAIL_COLLADA_MODELS="rail_collada_models"
RAIL_COLLADA_MODELS_GITHUB="git@github.com:WPI-RAIL/$RAIL_COLLADA_MODELS.git"

BUCKET="resources.robotwebtools.org"

# Grab current files
echo
echo "Downloading current meshes via GitHub..."
echo
rm -rf $TMP
mkdir $TMP
cd $TMP
git clone $RAIL_COLLADA_MODELS_GITHUB

# Copy non-compressed files
echo
echo "Copying non-compressed files..."
echo
mkdir -p "$TMP/$SYNC/$RAIL_COLLADA_MODELS/meshes/materials"
cp $TMP/$RAIL_COLLADA_MODELS/meshes/materials/* $TMP/$SYNC/$RAIL_COLLADA_MODELS/meshes/materials

# Compress the meshes
echo
echo "Compressing $RAIL_COLLADA_MODELS/meshes/*.min.dae..."
echo
cd $TMP/$RAIL_COLLADA_MODELS/meshes/
for f in *.min.dae ; do gzip -9 -c $f >> $TMP/$SYNC/$RAIL_COLLADA_MODELS/meshes/$f ; done

# Delete the old files
echo
echo "Removing s3://$BUCKET/$RAIL_COLLADA_MODELS..."
echo
aws s3 rm s3://$BUCKET/$RAIL_COLLADA_MODELS --recursive

# Upload new files
echo
echo "Uploading to s3://$BUCKET..."
echo
aws s3 cp $TMP/$SYNC s3://$BUCKET --recursive --content-encoding gzip --content-type application/xml --acl public-read --exclude "*.jpg"
aws s3 cp $TMP/$SYNC s3://$BUCKET --recursive --content-type image/jpeg --acl public-read --exclude "*.min.dae"

[rctoris]

Closing since the CARL demo models are now gzipped.