Cloud Source Blog
In This HP Cloud Source Blog, HP Expert, Christian Verstraete will examine cloud computing challenges, discuss practical approaches to cloud computing and suggest realistic solutions.

Cloud Computing & the Internet of Things

Source: freedigitalphotos.netThe world is increasingly becoming digital and interconnected. At the moment, our main focus is on human interactions, but that is changing quickly. The machine to machine communication is slowly but surely increasing. You will probably say… OK, but what does this have to do with me?

 

Well, let me run a little, somewhat futuristic, scenario. I’m walking down the street on a sunny afternoon. Suddenly I notice an ambulance, and to my astonishment, it’s stopping right beside me. A doctor jumps out and tells me, “Sir, please come in and lay down, we noticed you will have a heart attack in about 10 minutes and want to prevent that.”

Indeed, I have been carrying one of those health watches for a while and it has been transmitting some vital information to the cloud where algorithms have been monitoring the signals, looking for early signs of trouble.

 

You believe this is science fiction? Well, the technology exists as this wearable ECG patch demonstrates, so the above may happen sooner than you think.

 

Car electronics perform an ever increasing role in improving driving conditions. In car sensors and mathematical models adapt suspension, motor settings, etc., to speed, outside temperature and proximity of other cars. The next step is the connection of the car to its environment. Road and traffic conditions can then be provided to the car to further optimize the driving experience. In-road sensors can transmit road surface temperature and humidity for example, while anonymously tracking mobile phone locations and gives you throughput times on specific roads.

 

About a year ago, I talked to the provider of surveillance equipment about uploading camera images in the cloud where they could be analyzed for unusual behavior, calling the facility owner (or police) and streaming the video on his smart phone as soon as something strange is detected.

 

These are three examples of where interactions of sensors and their analysis in the cloud improves the way we live. Frankly, this is just two examples; creativity will be our limit here. But what is for sure is that such machine to machine interactions will only increase in the future. They will represent the largest part of big data in the near future.

 

Car manufacturers, health equipment providers, security system manufacturers and others have to start thinking how the digital world allows them to create new experiences for their users. Scalability and cost will obviously be key components in such offerings. So will manufacturers become service providers or will they consume services from external providers, using standard public cloud environments? That is a real good question. I believe they will start using the latter first, but when they realize the cost of doing so, they may look for more cost effective approaches.

 

Service characteristics

Look at any of the above examples. They all have in common the need for reasonably simple service components interacting together. The healthcare example is based around a signal analysis function, the car example focuses on mathematical models using sensor data to define key car settings (already used in Formula One and America Cup sailing and called two-way telemetry) and the surveillance one uses video image analysis.

 

In each, large information streams are coming in and being analyzed in real time. A decision on whether something has to be done is taken, and a process is started. If we assume action is only needed occasionally, we end-up with hundreds of thousands of instances of the analysis function and maybe some thousands of action processes.

 

A service oriented approach

You remember my blog entry a while ago on a service oriented world? We can implement the analysis function as a multi-tenant service that can quickly scale as demand grows. Depending on the outcome of the analysis service (action or no action), we can then trigger the appropriate process.

 

To make this extremely efficient, the analysis service needs to be optimized to the environment in which it will run. Typically in that resource pool only that analysis function will run, so we can really tweak the infrastructure and the service. This is precisely what HP Moonshot is out to do.

 

So, what is the scoop? HP Moonshot offers a brand new approach to infrastructure design, allowing you to use servers that are specialized for specific tasks. So, depending on what analysis service is used, we could choose a server cartridge (this is how a server that is plugged into a common chassis taking care of connections, communication, power and management) that best fits the task at hand. Potentially that server can be complemented with an FPGA, a GPU, an APU or a DSP. And this will work with 89% less energy, using 80% less space and 77% less cost. By tailoring the actual service to the functionality available in the server cartridge

 

With billions of sensors and devices linked to the cloud, being able to process them in real time with minimal infrastructure will be critical. And that is best done through “SaaS” type clouds designed to handle one type of workloads rather than many.

 

The raise of the specialized cloud

Today, clouds are general purpose environments using standard infrastructure to address a variety of workloads going from web site to high performance computing. Considering the wide range of the services, they are doing a good job. But frankly, is that the future? I don’t believe so. Any technology splinters over time into a number of directions that get more and more specialized. Take the bike for example; we started with a bike, now we have a running bike, a mountain bike, a city bike, a single speed, etc. Take the mobile phone; we started with a mobile phone, now we still have a standard, low cost mobile phone, but also a smart phone, specific phones for children or older people, etc. The same is doomed to happen with the cloud.

 

Now, this implies two things; first the cloud infrastructures will differentiate themselves increasingly in function of the task they focus on. And that is where HP Moonshot actually plays. Second, as a user, I will work in a hybrid cloud environment, where I will consume services from many clouds.

 

Such vision leaves a number of elements to be addressed. Let me name two. First, we will have to solve the issue of data and how I can integrate data across those different clouds. Second, we will need to access the end-to-end security across those different clouds. So we have plenty of areas where we can innovate. So, what are we waiting for? 

Labels: cloud| CloudSource
Leave a Comment

We encourage you to share your comments on this post. Comments are moderated and will be reviewed
and posted as promptly as possible during regular business hours

To ensure your comment is published, be sure to follow the community guidelines.

Be sure to enter a unique name. You can't reuse a name that's already in use.
Be sure to enter a unique email address. You can't reuse an email address that's already in use.
Type the characters you see in the picture above.Type the words you hear.
Search
Showing results for 
Search instead for 
Do you mean 
About the Author
Christian is responsible for building services focused on advising clients in their move to cloud, particularly from a business process and ...


Follow Us
Top Kudoed Posts
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation