Inside Three Intel Science and Technology Centers

An Interview with Jeff Parkhurst, Ph.D, Intel Labs

Jeff Parkhurst, Ph.D. Intel Labs

The Intel Science and Technology Center Program is a series of research collaborations that Intel is establishing with U.S. universities to identify and prototype revolutionary technology opportunities. The centers are designed to encourage closer collaboration among academic thought leaders in essential technology areas. 

Jeff Parkhurst is the Intel Program Director for Intel’s Science and Technology Centers for Big Data, Cloud Computing and Embedded Computing.  The three centers are working on new technology infrastructure essential for Big Data: a new ecosystem for Big Data.   

What do you think is the most pressing issue in computing today?

I think the biggest challenge is dealing with the huge amounts of data all around us. Data is being continuously generated by everything from smartphones and social media to sensor networks and instruments in hospitals, power plants, retail stores, city streets, and space. Technology has advanced to the point where we can capture the data, mine it and extract information from it. The challenge is effectively managing, parsing, and processing the data to extract useful, consumable information ― the gold nuggets.

We need more efficient and more effective ways not only to analyze and process data, but also to collect, store and share it ― ways that work together to keep up with the vastness, variety and velocity of today’s data.  A new ecosystem.

What are the essential components of the new ecosystem?

We need to get better at harvesting the data, arranging the data so that it can be parsed and processed, processing the data to extract information, and presenting the information for consumption.   We have three Intel Science and Technology Centers that are addressing various limitations in the current ecosystem:

  • The Big Data Center is focusing on new data management and analytics technologies for processing large quantities of data.
  • The Embedded Computing Center focuses on both harvesting and processing data locally on an embedded platform – for example, on a smartphone, a tablet, a scientific instrument or new device not yet invented.
  • The Cloud Computing Center is exploring how to efficiently and effectively provide resources to process data that cannot be processed locally.

Together, the new technologies coming out of these centers will help create essential tools to deal with this flood of new data.

“We also hope that the Big Data Center will have the additional impact of catalyzing the Big Data research community, drawing together world-class thought leaders to solve Big Data problems and providing them with more opportunities than they otherwise would have alone.” – Jeff Parkhurst, Intel Labs

What are the Centers working on?

The short answer is “a lot.” But here are a few examples:

Our Embedded Computing Center is working on making complex, computationally intensive algorithms work locally on smartphones and tablets. Today, our smartphones and tablets have cameras, sensors and displays, which make them great for collecting data.  But they aren’t good for processing data as the algorithms to do this were not meant to be run on a limited resource device. These devices don’t have the computational power and memory for activities like image recognition or video recognition for which these algorithms were originally developed. We expect the computational power of these devices to grow, but we also expect the resourcing requirements to grow commensurately.

For example, you can use your smartphone to take a photo of a landmark you spot on the street, but your phone cannot process the image to provide information (name, background information, etc.). To accomplish this, you would need to connect with a large data source (for context) and a more powerful platform (for analytics) to process and recognize the image.   As a second example, storing and searching data on a few hundred friends on your smartphone is possible; storing and searching data on billions of people, monuments, landmarks, etc., isn’t. At the same time, you don’t always want to be sending every piece of data back to a server farm to process, as there are bandwidth limitations and latency problems.

With the right new technologies for Embedded Computing Systems and the Cloud, we can create an ecosystem that’s intelligent enough to constantly balance the processing and analytics loads between local devices (e.g.,  smartphone, tablet) and remote processing like a server farm.

When we do this, it will be possible for your smartphone to not only identify landmarks, people and so on in a timely manner, but also provide suggestions on things you might want to see or do while you are in this neighborhood. This data, when processed, can provide information not only on what you want to know, but also what you need to know.

These same concepts can be applied to business, medicine, the sciences and other large data sets.  For example, a Boeing 747 creates 20 terabytes of data an hour.  A smartphone would not be able to process this much data, but with proper filtering and access to remote processing ― you could harvest a lot more of this data.

What about the Big Data Center?

The Big Data Center is working on new compute architectures and data management systems for Big Data, which we sorely need.  The vastness, velocity and variety of all this data has outstripped the capacity of today’s databases and systems.  The center is working on new database architectures – for example, array- and graph-oriented databases that offer faster, more practical ways to handle multi-dimensional data.   Traditional relational databases (RDBMS) are overwhelmed by today’s data . The Big Data Center is also working on new math and algorithms, new analytics, and faster ways to process streaming data.   And looking at such questions as how databases can use hardware more effectively to offload work and speed analysis.

One of the more important areas that our Big Data Center is working on is visualization.  Given this wealth of new information – the gold nuggets we can now mine continuously from Big Data – how can we best display it so that people can absorb the information and make decisions?

There are two challenges with displaying Big Data.  One, how do we provide it and in what context? Two, how do we summarize it ― for example in a picture or a video ― without losing important information in the process?  Think about a satellite map of the U.S., for example: it’s a high-definition picture with intense pixel density.  It doesn’t display well on a smartphone or tablet screen because those devices don’t have enough definition or pixels.  If you tried a pixel-to-pixel mapping, you’d have 100 pixels of data coming in for every pixel on the display. So, half your screen would be black.  You have to reduce the resolution to fit on your smartphone or tablet screen.  Or maybe not.

So, as you can see, with the help of our three centers, we’re working on the big challenges across the board ― from collecting and harvesting data, to loading, storing, analyzing and processing it, to displaying it.

How will Intel technology help create and advance these technologies?

Intel is investing significant resources ― in both dollars and people ― to each of these university centers and institutes to explore problems and solutions in these areas. This will provide us insight into the needs of tomorrow that will power the backbone for data processing: the processors for devices, video graphics, and server farms.

But we also realize that complete solutions must be explored in the form of new and innovative platforms in order for this entire ecosystem to function properly. This means our components must be very small and power-efficient to fit into an embedded platform where data is both harvested and consumed. We need to optimize our processors for the new kinds of computationally intense data management algorithms that I just talked about.  We need to provide new processing architectures that take advantage of  future disruptive memory technologies and fully exploit the potential for a hetero-core environment.

Intel is collaborating with university researchers worldwide in all these areas. Overall, we have initiated seven Intel Science and Technology Centers in the U.S. as well as five Intel Collaborative Research Institutes overseas.

Fast-forward to five years from now. How will these technologies that Intel is funding change the way computation works?

This is very-long-term research, and five years might be a little soon. However, ideally, research we are supporting would provide a means for processing vast quantities of data for analytic applications in medicine, scientific imagery, manufacturing, finance and so on.  Client devices like smartphones and tablets would be communicating directly with each other, and balancingl processing power between local and cloud resources.

Beyond the direct results of the research, we also hope that the Big Data Center will have the additional impact of catalyzing the Big Data research community, drawing together world-class thought leaders to solve Big Data problems and providing them with more opportunities than they otherwise would have alone. We would love to see increased momentum in Big Data technology innovation via new co-sponsors of the Big Data Center. Attracting more investors, startups, and researchers to the Big Data opportunity would be a welcome outcome to our initial investment.

Overall, we believe these efforts will lead to better information and a richer daily experience for everyone: consumers, business professionals, technical professionals, scientists, teachers and world leaders.

This entry was posted in Analytics, Big Data Architecture, Databases and Analytics, DBMS, ISTC for Big Data Blog, Visualizing Big Data and tagged , , , , . Bookmark the permalink.

Leave A Reply

Your email address will not be published. Required fields are marked *

9 − nine =