diff --git a/exercise.sh b/exercise.sh
index cca9a13..7703f40 100755
--- a/exercise.sh
+++ b/exercise.sh
@@ -1,214 +1,46 @@
 #!/usr/bin/env bash
 
-# **exercise.sh** - using the cloud can be fun
-
-# we will use the ``nova`` cli tool provided by the ``python-novaclient``
-# package
-#
-
-
-# This script exits on an error so that errors don't compound and you see
-# only the first error that occured.
-set -o errexit
-
-# Print the commands being run so that we can see the command that triggers
-# an error.  It is also useful for following allowing as the install occurs.
-set -o xtrace
-
-
-# Settings
-# ========
-
-# Use openrc + stackrc + localrc for settings
-source ./openrc
-
-# Get a token for clients that don't support service catalog
-# ==========================================================
-
-# manually create a token by querying keystone (sending JSON data).  Keystone
-# returns a token and catalog of endpoints.  We use python to parse the token
-# and save it.
-
-TOKEN=`curl -s -d  "{\"auth\":{\"passwordCredentials\": {\"username\": \"$NOVA_USERNAME\", \"password\": \"$NOVA_API_KEY\"}}}" -H "Content-type: application/json" http://$HOST_IP:5000/v2.0/tokens | python -c "import sys; import json; tok = json.loads(sys.stdin.read()); print tok['access']['token']['id'];"`
-
-# Launching a server
-# ==================
-
-# List servers for tenant:
-nova list
-
-# Images
-# ------
-
-# Nova has a **deprecated** way of listing images.
-nova image-list
-
-# But we recommend using glance directly
-glance -A $TOKEN index
-
-# Let's grab the id of the first AMI image to launch
-IMAGE=`glance -A $TOKEN index | egrep ami | cut -d" " -f1`
-
-# Security Groups
-# ---------------
-SECGROUP=test_secgroup
-
-# List of secgroups:
-nova secgroup-list
-
-# Create a secgroup
-nova secgroup-create $SECGROUP "test_secgroup description"
-
-# determine flavor
-# ----------------
-
-# List of flavors:
-nova flavor-list
-
-# and grab the first flavor in the list to launch
-FLAVOR=`nova flavor-list | head -n 4 | tail -n 1 | cut -d"|" -f2`
-
-NAME="myserver"
-
-nova boot --flavor $FLAVOR --image $IMAGE $NAME --security_groups=$SECGROUP
-
-# Testing
-# =======
-
-# First check if it spins up (becomes active and responds to ping on
-# internal ip).  If you run this script from a nova node, you should
-# bypass security groups and have direct access to the server.
-
-# Waiting for boot
-# ----------------
-
-# Max time to wait while vm goes from build to active state
-ACTIVE_TIMEOUT=${ACTIVE_TIMEOUT:-10}
-
-# Max time till the vm is bootable
-BOOT_TIMEOUT=${BOOT_TIMEOUT:-15}
-
-# Max time to wait for proper association and dis-association.
-ASSOCIATE_TIMEOUT=${ASSOCIATE_TIMEOUT:-10}
-
-# check that the status is active within ACTIVE_TIMEOUT seconds
-if ! timeout $ACTIVE_TIMEOUT sh -c "while ! nova show $NAME | grep status | grep -q ACTIVE; do sleep 1; done"; then
-    echo "server didn't become active!"
-    exit 1
-fi
-
-# get the IP of the server
-IP=`nova show $NAME | grep "private network" | cut -d"|" -f3`
-
-# for single node deployments, we can ping private ips
-MULTI_HOST=${MULTI_HOST:-0}
-if [ "$MULTI_HOST" = "0" ]; then
-    # sometimes the first ping fails (10 seconds isn't enough time for the VM's
-    # network to respond?), so let's ping for a default of 15 seconds with a
-    # timeout of a second for each ping.
-    if ! timeout $BOOT_TIMEOUT sh -c "while ! ping -c1 -w1 $IP; do sleep 1; done"; then
-        echo "Couldn't ping server"
-        exit 1
+# Run everything in the exercises/ directory that isn't explicitly disabled
+
+# comma separated list of script basenames to skip
+# to refrain from exercising euca.sh use SKIP_EXERCISES=euca
+SKIP_EXERCISES=${SKIP_EXERCISES:-""}
+
+# Locate the scripts we should run
+EXERCISE_DIR=$(dirname "$0")/exercises
+basenames=$(for b in `ls $EXERCISE_DIR/*.sh`; do basename $b .sh; done)
+
+# Track the state of each script
+passes=""
+failures=""
+skips=""
+
+# Loop over each possible script (by basename)
+for script in $basenames; do
+    if [[ "$SKIP_EXERCISES" =~ $script ]] ; then
+        skips="$skips $script"
+    else
+        echo =========================
+        echo Running $script
+        echo =========================
+        $EXERCISE_DIR/$script.sh
+        if [[ $? -ne 0 ]] ; then
+            failures="$failures $script"
+        else
+            passes="$passes $script"
+        fi
     fi
-else
-    # On a multi-host system, without vm net access, do a sleep to wait for the boot
-    sleep $BOOT_TIMEOUT
-fi
-
-# Security Groups & Floating IPs
-# ------------------------------
-
-# allow icmp traffic (ping)
-nova secgroup-add-rule $SECGROUP icmp -1 -1 0.0.0.0/0
-
-# List rules for a secgroup
-nova secgroup-list-rules $SECGROUP
-
-# allocate a floating ip
-nova floating-ip-create
-
-# store  floating address
-FLOATING_IP=`nova floating-ip-list | grep None | head -1 | cut -d '|' -f2 | sed 's/ //g'`
-
-# add floating ip to our server
-nova add-floating-ip $NAME $FLOATING_IP
-
-# test we can ping our floating ip within ASSOCIATE_TIMEOUT seconds
-if ! timeout $ASSOCIATE_TIMEOUT sh -c "while ! ping -c1 -w1 $FLOATING_IP; do sleep 1; done"; then
-    echo "Couldn't ping server with floating ip"
-    exit 1
-fi
-
-# pause the VM and verify we can't ping it anymore
-nova pause $NAME
-
-sleep 2
-
-if ( ping -c1 -w1 $IP); then
-    echo "Pause failure - ping shouldn't work"
-    exit 1
-fi
-
-if ( ping -c1 -w1 $FLOATING_IP); then
-    echo "Pause failure - ping floating ips shouldn't work"
-    exit 1
-fi
-
-# unpause the VM and verify we can ping it again
-nova unpause $NAME
-
-sleep 2
-
-ping -c1 -w1 $IP
-
-# dis-allow icmp traffic (ping)
-nova secgroup-delete-rule $SECGROUP icmp -1 -1 0.0.0.0/0
-
-# FIXME (anthony): make xs support security groups
-if [ "$VIRT_DRIVER" != "xenserver" ]; then
-    # test we can aren't able to ping our floating ip within ASSOCIATE_TIMEOUT seconds
-    if ! timeout $ASSOCIATE_TIMEOUT sh -c "while ping -c1 -w1 $FLOATING_IP; do sleep 1; done"; then
-        print "Security group failure - ping should not be allowed!"
-        echo "Couldn't ping server with floating ip"
-        exit 1
-    fi
-fi
-
-# de-allocate the floating ip
-nova floating-ip-delete $FLOATING_IP
-
-# shutdown the server
-nova delete $NAME
-
-# Delete a secgroup
-nova secgroup-delete $SECGROUP
-
-# FIXME: validate shutdown within 5 seconds
-# (nova show $NAME returns 1 or status != ACTIVE)?
-
-# Testing Euca2ools
-# ==================
-
-# make sure that we can describe instances
-euca-describe-instances
-
-if [[ "$ENABLED_SERVICES" =~ "swift" ]]; then
-    # Testing Swift
-    # =============
-
-    # Check if we have to swift via keystone
-    swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD stat
-
-    # We start by creating a test container
-    swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD post testcontainer
-
-    # add some files into it.
-    swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD upload testcontainer /etc/issue
-
-    # list them
-    swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD list testcontainer 
-
-    # And we may want to delete them now that we have tested that
-    # everything works.
-    swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD delete testcontainer
-fi
+done
+
+# output status of exercise run
+echo =========================
+echo =========================
+for script in $skips; do
+    echo SKIP $script
+done
+for script in $passes; do
+    echo PASS $script
+done
+for script in $failures; do
+    echo FAILED $script
+done
diff --git a/exercises/euca.sh b/exercises/euca.sh
new file mode 100755
index 0000000..9605ace
--- /dev/null
+++ b/exercises/euca.sh
@@ -0,0 +1,36 @@
+#!/usr/bin/env bash
+
+# we will use the ``euca2ools`` cli tool that wraps the python boto 
+# library to test ec2 compatibility
+#
+
+# This script exits on an error so that errors don't compound and you see
+# only the first error that occured.
+set -o errexit
+
+# Print the commands being run so that we can see the command that triggers
+# an error.  It is also useful for following allowing as the install occurs.
+set -o xtrace
+
+
+# Settings
+# ========
+
+# Use openrc + stackrc + localrc for settings
+pushd $(cd $(dirname "$0")/.. && pwd)
+source ./openrc
+popd
+
+# find a machine image to boot
+IMAGE=`euca-describe-images | grep machine | cut -f2`
+
+# launch it
+INSTANCE=`euca-run-instances $IMAGE | grep INSTANCE | cut -f2`
+
+# assure it has booted within a reasonable time
+if ! timeout $RUNNING_TIMEOUT sh -c "while euca-describe-instances $INSTANCE | grep -q running; do sleep 1; done"; then
+    echo "server didn't become active within $RUNNING_TIMEOUT seconds"
+    exit 1
+fi
+
+euca-terminate-instances $INSTANCE
diff --git a/exercises/floating_ips.sh b/exercises/floating_ips.sh
new file mode 100755
index 0000000..75046d1
--- /dev/null
+++ b/exercises/floating_ips.sh
@@ -0,0 +1,190 @@
+#!/usr/bin/env bash
+
+# **exercise.sh** - using the cloud can be fun
+
+# we will use the ``nova`` cli tool provided by the ``python-novaclient``
+# package
+#
+
+
+# This script exits on an error so that errors don't compound and you see
+# only the first error that occured.
+set -o errexit
+
+# Print the commands being run so that we can see the command that triggers
+# an error.  It is also useful for following allowing as the install occurs.
+set -o xtrace
+
+
+# Settings
+# ========
+
+# Use openrc + stackrc + localrc for settings
+pushd $(cd $(dirname "$0")/.. && pwd)
+source ./openrc
+popd
+
+# Get a token for clients that don't support service catalog
+# ==========================================================
+
+# manually create a token by querying keystone (sending JSON data).  Keystone
+# returns a token and catalog of endpoints.  We use python to parse the token
+# and save it.
+
+TOKEN=`curl -s -d  "{\"auth\":{\"passwordCredentials\": {\"username\": \"$NOVA_USERNAME\", \"password\": \"$NOVA_API_KEY\"}}}" -H "Content-type: application/json" http://$HOST_IP:5000/v2.0/tokens | python -c "import sys; import json; tok = json.loads(sys.stdin.read()); print tok['access']['token']['id'];"`
+
+# Launching a server
+# ==================
+
+# List servers for tenant:
+nova list
+
+# Images
+# ------
+
+# Nova has a **deprecated** way of listing images.
+nova image-list
+
+# But we recommend using glance directly
+glance -A $TOKEN index
+
+# Let's grab the id of the first AMI image to launch
+IMAGE=`glance -A $TOKEN index | egrep ami | cut -d" " -f1`
+
+# Security Groups
+# ---------------
+SECGROUP=test_secgroup
+
+# List of secgroups:
+nova secgroup-list
+
+# Create a secgroup
+nova secgroup-create $SECGROUP "test_secgroup description"
+
+# determine flavor
+# ----------------
+
+# List of flavors:
+nova flavor-list
+
+# and grab the first flavor in the list to launch
+FLAVOR=`nova flavor-list | head -n 4 | tail -n 1 | cut -d"|" -f2`
+
+NAME="myserver"
+
+nova boot --flavor $FLAVOR --image $IMAGE $NAME --security_groups=$SECGROUP
+
+# Testing
+# =======
+
+# First check if it spins up (becomes active and responds to ping on
+# internal ip).  If you run this script from a nova node, you should
+# bypass security groups and have direct access to the server.
+
+# Waiting for boot
+# ----------------
+
+# Max time to wait while vm goes from build to active state
+ACTIVE_TIMEOUT=${ACTIVE_TIMEOUT:-10}
+
+# Max time till the vm is bootable
+BOOT_TIMEOUT=${BOOT_TIMEOUT:-15}
+
+# Max time to wait for proper association and dis-association.
+ASSOCIATE_TIMEOUT=${ASSOCIATE_TIMEOUT:-10}
+
+# check that the status is active within ACTIVE_TIMEOUT seconds
+if ! timeout $ACTIVE_TIMEOUT sh -c "while ! nova show $NAME | grep status | grep -q ACTIVE; do sleep 1; done"; then
+    echo "server didn't become active!"
+    exit 1
+fi
+
+# get the IP of the server
+IP=`nova show $NAME | grep "private network" | cut -d"|" -f3`
+
+# for single node deployments, we can ping private ips
+MULTI_HOST=${MULTI_HOST:-0}
+if [ "$MULTI_HOST" = "0" ]; then
+    # sometimes the first ping fails (10 seconds isn't enough time for the VM's
+    # network to respond?), so let's ping for a default of 15 seconds with a
+    # timeout of a second for each ping.
+    if ! timeout $BOOT_TIMEOUT sh -c "while ! ping -c1 -w1 $IP; do sleep 1; done"; then
+        echo "Couldn't ping server"
+        exit 1
+    fi
+else
+    # On a multi-host system, without vm net access, do a sleep to wait for the boot
+    sleep $BOOT_TIMEOUT
+fi
+
+# Security Groups & Floating IPs
+# ------------------------------
+
+# allow icmp traffic (ping)
+nova secgroup-add-rule $SECGROUP icmp -1 -1 0.0.0.0/0
+
+# List rules for a secgroup
+nova secgroup-list-rules $SECGROUP
+
+# allocate a floating ip
+nova floating-ip-create
+
+# store  floating address
+FLOATING_IP=`nova floating-ip-list | grep None | head -1 | cut -d '|' -f2 | sed 's/ //g'`
+
+# add floating ip to our server
+nova add-floating-ip $NAME $FLOATING_IP
+
+# test we can ping our floating ip within ASSOCIATE_TIMEOUT seconds
+if ! timeout $ASSOCIATE_TIMEOUT sh -c "while ! ping -c1 -w1 $FLOATING_IP; do sleep 1; done"; then
+    echo "Couldn't ping server with floating ip"
+    exit 1
+fi
+
+# pause the VM and verify we can't ping it anymore
+nova pause $NAME
+
+sleep 2
+
+if ( ping -c1 -w1 $IP); then
+    echo "Pause failure - ping shouldn't work"
+    exit 1
+fi
+
+if ( ping -c1 -w1 $FLOATING_IP); then
+    echo "Pause failure - ping floating ips shouldn't work"
+    exit 1
+fi
+
+# unpause the VM and verify we can ping it again
+nova unpause $NAME
+
+sleep 2
+
+ping -c1 -w1 $IP
+
+# dis-allow icmp traffic (ping)
+nova secgroup-delete-rule $SECGROUP icmp -1 -1 0.0.0.0/0
+
+# FIXME (anthony): make xs support security groups
+if [ "$VIRT_DRIVER" != "xenserver" ]; then
+    # test we can aren't able to ping our floating ip within ASSOCIATE_TIMEOUT seconds
+    if ! timeout $ASSOCIATE_TIMEOUT sh -c "while ping -c1 -w1 $FLOATING_IP; do sleep 1; done"; then
+        print "Security group failure - ping should not be allowed!"
+        echo "Couldn't ping server with floating ip"
+        exit 1
+    fi
+fi
+
+# de-allocate the floating ip
+nova floating-ip-delete $FLOATING_IP
+
+# shutdown the server
+nova delete $NAME
+
+# Delete a secgroup
+nova secgroup-delete $SECGROUP
+
+# FIXME: validate shutdown within 5 seconds
+# (nova show $NAME returns 1 or status != ACTIVE)?
+
diff --git a/exercises/swift.sh b/exercises/swift.sh
new file mode 100755
index 0000000..f7be099
--- /dev/null
+++ b/exercises/swift.sh
@@ -0,0 +1,40 @@
+#!/usr/bin/env bash
+
+# Test swift via the command line tools that ship with it.
+
+# This script exits on an error so that errors don't compound and you see
+# only the first error that occured.
+set -o errexit
+
+# Print the commands being run so that we can see the command that triggers
+# an error.  It is also useful for following allowing as the install occurs.
+set -o xtrace
+
+
+# Settings
+# ========
+
+# Use openrc + stackrc + localrc for settings
+pushd $(cd $(dirname "$0")/.. && pwd)
+source ./openrc
+popd
+
+
+# Testing Swift
+# =============
+
+# Check if we have to swift via keystone
+swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD stat
+
+# We start by creating a test container
+swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD post testcontainer
+
+# add some files into it.
+swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD upload testcontainer /etc/issue
+
+# list them
+swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD list testcontainer
+
+# And we may want to delete them now that we have tested that
+# everything works.
+swift --auth-version 2 -A http://${HOST_IP}:5000/v2.0 -U admin -K $ADMIN_PASSWORD delete testcontainer
diff --git a/openrc b/openrc
index 324780b..4b36112 100644
--- a/openrc
+++ b/openrc
@@ -49,3 +49,14 @@ export EC2_SECRET_KEY=${ADMIN_PASSWORD:-secrete}
 # set log level to DEBUG (helps debug issues)
 # export NOVACLIENT_DEBUG=1
 
+# Max time till the vm is bootable
+export BOOT_TIMEOUT=${BOOT_TIMEOUT:-15}
+
+# Max time to wait while vm goes from build to active state
+export ACTIVE_TIMEOUT=${ACTIVE_TIMEOUT:-10}
+
+# Max time from run instance command until it is running
+export RUNNING_TIMEOUT=${RUNNING_TIMEOUT:-$(($BOOT_TIMEOUT + $ACTIVE_TIMEOUT))}
+
+# Max time to wait for proper IP association and dis-association.
+export ASSOCIATE_TIMEOUT=${ASSOCIATE_TIMEOUT:-10}