Meeting Prof. Dr. Dan Reed

Damir Dobric Posts

Next talks:



Follow me on Twitter: #ddobric




Frankfurt am Main – Germany, October 28 – 29, 2010
International Supercomputing Conference 2010 has been hosted this year in Frankfurt am Main - Germany. ISC Cloud'10 is a conference that focuses on massive data processing and large application simulations in the Cloud.This inaugural Cloud program is organized by an international committee of recognized experts in areas such as cloud computing, high performance computing, and data-and compute-intensive applications in the cloud, in industry and in research.

The number of internationally recognized speakers have been invited this year to give an overview of the state of research in this area. For example Dan Reed from Microsoft with a keynote talk on Technical Clouds: Seeding Discovery and Kathy Yelick from Lawrence Berkeley Laboratory with her keynote on Science in the Clouds: A View from Berkeley.

I had a pleasure to meet in person Prof.Dr. Dan Reed who practically started the conference with a great keynote. Dan is by the way Microsoft Corporate Vice President, Technology Strategy and Policy & Extreme Computing Group, Microsoft, USA.
He is working on many interesting projects, which could possible change the future of information technology or even already has changed it.I have tried to recap our conversation in few major topics.

The Society, new technologies and the Cloud

At the moment, I’m disappointed about how we (people in Germany) are ready to adopt new ideas. For example Google-Street View. Too many people are too scared. Are we getting too old that fast? because of this, I asked following question:

Dan, what is easier? To develop a new technology or to adopt it by society?

Dan thinks that we are at the moment just renaming more or less already existing terms and technologies. For example, many years ago we started to talk about “hosted services”. Isn’t that somehow SasS?
It seems to be just a question of time when things like Windows Azure will be totally adopted. In research space people work with high performance computing. In application development space people use more domain specific tools and do not need raw parallel computing.

When things like massive scale in Windows Azure will be adopted is also a measure of hardware and performance. For example, do I need to calculate my Excel sheets on my single machine or to span it across multiple machines in the cloud?

In third, consumers space, the cloud has already begun by integrating of client devices, laptops, mobile phones etc.

It will take a time, by the cloud is adopting by society step by step.

Security consideration as “Made in Germany”

Because I have a feeling that society in Germany seems to be surprisingly sometimes very conservative and in some points often even destructive, when new technologies need be adopted. My colleges in US are adopting the cloud more quickly than my colleges in Germany.
I asked why:
Dan, What is more safe? My credit card in the pocked or my data in the cloud?

Dan found this question very, amusing. :)

He answered shortly: ”The security is a perception”. It is some kind of  social trust to some person”. Dan thinks, that perception of security in the cloud can be achieved by adopting of common international policy. It is a conflict between economy and technology. One of major problems are different cultural expectations and policies. We need to look for commonalities.
In technology we need approvable chain of trust, which would involve different penetrations. We will also need a trusted platform models with cryptographic support. This kind of security should enable external users to verify that what they expect to see.
One of very important researching in this context is for example “searchable encryption”. Soon we will need to run computation on encrypted data in general.
Such kind of technology will definitely have an impact on perception of security and trust in the cloud.

Cooperation between Extreme Computing research and Windows Azure Team

Because I have often a feeling that some research projects sometimes are a bit disconnected from real industry needs, I just had to ask this question:

Dan, how tightly do you work with Windows Azure team?

Dan says, that lot of research work is already consisted part of Windows Azure.
I find also few next statements very interesting for community.
At the moment Dan’s group is working on the next generation of hardware for cloud. For example, they are trying to build the processor, computer etc. to spent as less energy as possible.

About the next big thing

Because I’m working in highly innovative software company (daenet) I just have to know what his group is doing right now?

Dan, What is your most exciting project at the moment which is not under NDA?

Again, he found my question funny.
As I understood, the major requirement in his group is at the moment to build the hardware which is able  reuse energy like solar and wind energy sources . The structure and concepts of such hardware will have a strong impact on infrastructure and implicitly also on and programming model and applications itself . In this concept things like power grids,  solar and wind energy play a very important role.
It is lot of about eco-friendly and much more smaller datacenters. More efficient cooling, by complete redesign of hardware. And, it is not all about technology. Much more other things come into play. For example energy cost or even geological and political stability.

In this context I also asked following question:

Is there some even more exciting project, which is a top secret one?

Dan laughed again and answered “I couldn't’ tell you that”. :)

The message for developers

I was interested to know when we developers need to move to Windows Azure programming model?
He thinks that we all are today in a division of the future. In fact, we need to build software as usual today. But, the Cloud is inside of Microsoft truly more than just a marketing story. Microsoft is committed to Cloud.
I asked him to define more precisely the time when we will no more build any other than cloud capable software. (please all of you who is reading this,  do not understand this absolutely). I wanted to know does this point of time falls in this decade?
He said, “Definitely.It is sooner as we think”.

ISC '11

Dan, thank you for your time.


Posted Oct 31 2010, 09:28 PM by Damir Dobric is a .Net Community Blog powered by daenet GmbH.