Monitor water in 3 steps by SensePlanet:   http://sensepla.net

Furthermore, we advocate for leveraging social web to unleash the “Social Grid”. The concept is simple: make things, more specifically sensors, communicate with social networks.

For that purpose, a RESTFul service has been prototyped so as to convey sensor measurements to social networks. In the coming weeks, I will document the design, then the API. The code will be made available.
As of now, only status updates and messaging are supported for Facebook and Twitter but in the future, this service will be a platform for rich social applications.

Today, I’d like to unveil the back-end administration page ( in STRUTS..I know, it is old but heck, it works! ) which interfaces with the core web services ( Tomcat + MySQL ).

It is here: http://socialgrid.senseplanet.com .
For info, “lokicool” is the first name given to this project.

Let me know what you guys think! ( Besides that this post lead image is lame! )

Lately, we’ve been working on a data aggregating services that allows our wireless sensors to communicate easily with social networks such as Facebook and Twitter ( similar to a service like Ping.fm, but for sensors)
This prototyping has led to a modular design to support future social networks as well as other output streams ( messaging, RSS, … )
For that matter, the plugin architecture with a RESTFul API  has been deemed to be the most relevant. Please refer below for a high-level design:

socialEnergy: a service design

By relying on this architecture, we could tie each sensor to a list of social networks with which to interact.
The unique identifier to differentiate each sensor is the MAC address.
The web services client should be running either on the sensor or ported out on the smart home gateway ( refer to previous post ).
The plugin interface translates the data coming from the sensor into 3rd party API calls.
The work queue can be simply a Java messaging system.

Overall, this design enables social capabilities to sensor network. More importantly, the system can roll out new plugins without impacting the software running on the sensor side.

NOTE: You can see how the admin page of our server looks like here: socialgrid.senseplanet.com

Meshlium

As I mentioned in my previous post, the home router will most likely become the smart home gateway by bridging things ( sensors, appliances,..) and the Internet wirelessly.

The home computer, which holds this role as of now, might be just used for wired connectivity ( camera, high-speed high-drive, …) and web browsing.

For that matter, Libelium has unveiled “the first multiprotocol mesh router to combine Wifi, ZigBee, GPRS, Bluetooth and GPS technologies. It can also be powered by solar panels, or connected to the power supply or car lighter.” The name if this beast: Meshlium.

It is a breakthrough in the “web of things” industry. Not only, Meshlium hardware supports almost all the LAN+cellular technologies but the management software will provide the harness to implement innovative pervasive services in a breeze.
Not to mention that the GPRS connectivity will not tie wireless sensors to a home/DSL access anymore. Eventually, it should foster rural and wide-scale use.

What else beside that you should definitely check it out. It is just awesome!

WHAT IS AMEE?

While sensors become more and more ubiquitous, the question of data aggregation is raised as the next step for standardization.
Should the industry go with proprietary solutions to aggregate data from different sources or with a standardized approach.
With regards to the latter, AMEE leads the march by providing a protocol and a database to store data from sensors. The protocol is simply and smartly a REST API.

WHY IS AMEE SWEET?

Not only it will allow sensors and smart meters companies to focus on their core expertise, but it will also provide a great tool for business intelligence and environmental institution to gather and analyze data from heterogenous sources such as water/electrical consumption, pollution,…
Somehow, AMEE can  be deemed as the window on the smart grid.

Despite high expectations, I have some concerns about security and reliability which should be clarified and emphasized to get more traction

I recommend reading through their high-level presentation.
More specifically, have a look about how IBM uses this distributed data warehouse for a home energy monitoring project.

HOW TO USE AMEE?

AMEE is a REST API which can be accessed after signing up for a key.
The admin portal can then be accessed with the credentials sent over to your email inbox.

The admin portal will provide enough information to get started: json format that is.

All in all, I would advocate for using this tool for environmental project. Not only it will remove the burden of hosting a database to store your data but it will allow you to share your data and participate in large-scale studies.

Nagios is a IT monitoring solution used widely in network data center. By supplementing their solution with environmental monitoring, greener data center is just one wire away from reality.
It really provides an easy upgrade for network engineers to control their energy consumption, hence their environment footprint.

I wonder how widely the Nagios approach will be adopted, as opposed to that taken by Sentilla , which relies on wireless mesh sensors.

Broadband internet access is, albeit widespread,  very expensive in term of infrastructure, the last mile still being not easy to bridge.

While the demand for ubiquitous broadband grows tremendously, Wifi mesh network looks like the best solution to keep up the pace. Indeed, Wifi mesh network deployment is a breeze : communities, hobbyists and municipalities should definitely take a look at this technology to spread internet access everywhere.

More specifically, the “state of the art” WIFI mesh network consists of a bunch of WIFI routers  along with a service to manage them remotely ( for billing and network configuration)
A mesh router can act like a gateway or a repeater. The router acting like a gateway is connected to the internet whereas the repeater just conveys the data to the gateway through the best route. That way, a network coverage can be widen just by adding repeaters. The advantages of this mesh solution can be summarized as below:

• low-cost routers: from 49$ up to 900$ for solar powered.
• scalability: network coverage can be improved by adding additional nodes
• reliability: if a node is not functional, another one will take over so as to convey data through another path.
• easy management: server is available to configure and manage the routers, hence the network remotely.

I’ve been buying this technology from  two providers with different vision: Meraki and Open-Mesh.

Both Meraki and Open-Mesh provide the routers and the management solution.
Meraki has been my choice for a while since there hosted management services provides everything a network admin needs ( billing, configuration, update, stats,..). The Meraki routers are also reliable and beautiful.
However, Meraki has increased their price and their cheapest router is at 149$ ( compared to 50$ before !). Also, the routers are tied to their hosted management service such that the network admin has no choice as to which software runs in the routers. That’s why I’ve opted lately for the Open-Mesh solution which is a no surprise offer. Also, the network admin can host the management server thanks to OrangeMesh, an open source initiative.

As a conclusion, one should look at the open-mesh products especially to lower cost and have a better control over their networks.
Meraki is still the  best if one wants a no-brainer solution ( but 3 times more expensive than their open-source counterpart )

I love Paris: a city packed with history, culture, elegance as well as cranky people and traffic jam!….When I come back to Paris from my new hometown San Francisco, I am just flabbergasted by the streets jam-packed with cars and trucks…all day long. Willy nilly, it has to be fixed.

And here comes a ray of hope:  Sensys Networks with their smart traffic management solution based on wireless sensors can help reduce traffic congestion. Indeed, “if you cannot measure it, you cannot improve” statement is also very true for traffic management and Sensys Networks with their magnetic, 10 year battery life, wireless powered sensors can help municipalities glean real-time traffic information so as to improve their transportation grid.

This is crucial: in cities teeming with cars and traffic, technology like Sensys Networks is the key towards greener and monitorable neighborhoods for us, human, to live in.

You can thumb through a good review here.

Indeed, I’ve been working on a RESTful API for my wireless sensors. I was a newbie and despite many roadblocks, let me assure you that it is amazing how server/web services technologies are accessible nowadays!
Anyway, now that I have a home-brew Java web services running locally, I tried to load it on a cloud server and below is the procedure I followed.

Cloud server providers:

I’ve been looking at Amazon EC2 and Mosso. I’ve decided to go for the latter:

• Price: Mosso provides cloud server starting at about 15$ a month for 256MB RAM or 30$ for a 512MB.
• Ease-of-use: Creating a server instance is straight-forward with Mosso.
  Once the server is created, server informations ( login/pw/ip ) are sent over by email. SSH connection gives you root permissions.
  Whereas EC2 server configuration is more tricky
• Reliability, scalability, security: I would discuss those topics later on since I haven’t been confronted to them yet.
  Should you have any opinion, please chime in!

Server creation with Mosso:

Launching a server instance with Mosso is a breeze. Once an account is registered, one can access a neat admin page whereby a new server instance can be created on-demand.
Several OS are available and I’ve picked up CentOS 5.2 because it was the first one in the list !
Below is a server creation page screenshot:

Mosso Cloud Server

MySQL installation:

To install MySQL, just follow the procedure below:

1. Connect to your server:
ssh root@yourIP
2. Download and install MySQL thanks to ‘yum” command:
yum install mysql-server mysql mysql-devel
3. Start MySQL:
/etc/init.d/mysqld start
4. Add MySQL user:
CREATE USER 'yourUsername' IDENTIFIED BY 'yourPassword';
CREATE USER 'yourUsername'@'localhost' IDENTIFIED BY 'yourPassword';
GRANT ALL ON *.* TO 'yourUsername' IDENTIFIED BY 'yourPassword';
GRANT ALL ON *.* TO 'yourUsername'@'localhost' IDENTIFIED BY 'yourPassword';
5. That’s it!

Tomcat 6, Java 1.6 and ActiveMQ 5.2:

Although ‘yum’ could be used,  I’d rather download locally the packages for Tomact, Java and ActiveMQ.
Now, let’s schlep those packages to my cloud server:

1. Download Tomcat 6 , Java 6 and ActiveMQ 5.2
2. Uncompress and copy the binaries respectively under /usr/tomcat ,  /usr/java/ and usr/activemq/ as below:
scp -r /jdk1.6.0_05 root@yourIP:/usr/java/
scp -r 6.0.18/ root@yourIP:/usr/tomcat/
scp -r activemq5.2/ root@yourIP:/usr/activemq

Environment variables:

Let’s create a .bash_profile file to set environment variables properly as below:
My .bash profile:

#java
export JAVA_HOME=/usr/java/jdk1.6.0_05/
export JRE_HOME=/usr/java/jdk1.6.0_05/jre/
#tomcat
export CATALINA_HOME=/usr/tomcat/6.0.18/
#activemq
export ACTIVEMQ_HOME=/usr/activemq/5.2/
#mysql
export MYSQL=/usr/bin/
export mysql=/usr/bin/mysql
export mysqladmin=/usr/bin/mysqladmin
#export PATH
export PATH=$JAVA_HOME/bin:$CATALINA_HOME/bin:$JRE_HOME/bin:$ACTIVEMQ_HOME:$MYSQL:$PATH

Start Tomcat and ActiveMQ:

1. Start ActiveMQ ( if needed ):
/usr/activemq/5.2/bin/activemq
2. Start Tomcat:
/usr/tomcat/6.0.18/bin/startup.sh
3. That’s it…. check whether http://yourIP:8080 loads the Tomcat admin page.

Conclusion:

Within few steps away, one can easily create a server instance on-demand.
Mosso Cloud Server is very appealing and user-friendly. However  I am contemplating to switch over to Amazon EC2 for evaluation purposes.
The reasoning is that Mosso does not provide a way to save home-brew server configurations with specific load of applications whereas Amazon EC2 with its AMIs ( Amazon Machine Instance ) may provide this useful capability.

I’ll keep you posted!