Arizona Technology and Innovation: Jetstream Cyber-Platform

Dec. 3, 2014

Transcript:

Ted Simons: Tonight's look at Arizona technology and innovation focuses on how scientists and engineers can store, share, and analyze data through jetstream, a new all-science Cyber computing platform. A University of Arizona-based program known as the I-plant collaborative will play an important role in the development and operation of jetstream, here to tell us more is Nirav Merchant, the co-principal and investigator of the I-plant collaborative and director of biocomputing at Arizona research labs and Edwin skidmore, the I-planted's assistant director of infrastructure and head of the u of a's cloud computing platform known as atmosphere. What is jet stream?

Nirav Merchant: It is a specialized service targeted for scientist and is engineers to make it easy for them to make use of the massive cloud computing systems, and it is an interface that makes them productive with the analysis.

Ted Simons: How does it differ from what's available or what was available prior?

Nirav Merchant: So prior everything you have seen so far has been capacity-based. It's big, it's large but it lacks the interfaces to use them right away. So, you get the capacity, but you spend more time trying to figure it out. This is something that you can use right away, and you have your applications and data, and it's all there and connected.

Ted Simons: And is this entirely cloud-based? How much is cloud-based here?

Edwin Skidmore: So, atmosphere can integrate with multiple different clouds, so for jet stream, there will be two different cloud installations, and atmosphere will connect both of those. So, 100% will be clouds.

Ted Simons: So it's cloud-to-cloud information, in other words?

Edwin Skidmore: Yes.

Ted Simons: How did all this get started? Talk to us about the genesis of all of this.

Edwin Skidmore: Atmosphere was developed through the project, the I-plant collaborative. Early on in the project, we realized the need for scientists to have access to highly customizable, shareable, virtual machines, virtual environments, and so out came from that was the sphere. Very simple to use web portal to connect the scientists.

Ted Simons: And this is for scientists, only. This is dedicated to scientists?

Edwin Skidmore: So, scientists and, actually, educators who want to teach the next generation of scientists on how to do computing on the cloud.

Ted Simons: Give me more information on the I-plant collaborative.

Nirav Merchant: So the I-plant was established in 2008, and it was dedicated to create new infrastructure for scientists to address the next generation analytics, and after the first five years, the national science foundation, which invested 50 million dollars, realized what we had built was so generalizable that it was useful for the life science as a whole discipline, and so for the next five years, they have asked us to support animal sciences and anybody in life sciences that needs to work with that data.

Ted Simons: So as far as I-plant involves the jet stream project again, it's, basically, one and two is next and we go on to three in the future. Is that what -- it is a step-by-step process?

Nirav Merchant: Absolutely. This was something that was born here and is now national, and it is going to go international because we have two other projects that are funded outside the U.S. based on this work. So, it is really growing.

Ted Simons: Ok. It's growing. Why is it important? Talk about Cyber infrastructure and what is needed as far as the sciences are concerned.

Edwin Skidmore: Well, scientific discovery really is, in particular, using computational tools. There is really a need for scientists to have the access, convenient access to compute and large storage, and particularly in this day and age, where there is large amounts of data and complex work flows to do analysis, atmosphere and jet stream will fit that need. Basically, you know, making it easier for scientists to use the cloud and not have to worry about the details of what infrastructure, what goes into building an infrastructure.

Ted Simons: We are talking huge data sets and complex analysis all easier to get around?

Edwin Skidmore: Easier and shareable. So, a scientist may develop a very complex work flow, and now they can share conveniently with other scientists and engineers.

Ted Simons: It sounds like we have more democratizing going on here.

Nirav Merchant: You used the word we like a lot, which is data and access, earlier, the groups that had the machines, were the ones that were at the forefront, and now, anybody with a good idea, a jet stream is going to be free access. It does not cost you any money as a scientist. You write a proposal. You get time on it. If you have a good idea, you can work with it. You don't have to really worry about owning a big machine to do that. You are very right.

Ted Simons: You have seen what atmosphere, what you are going to see with the jet stream. A better indication for those who are not involved in science and research on just what a difference we're talking about here.

Nirav Merchant: So, the difference between the jet stream --

Ted Simons: No, the difference between before atmosphere and jet stream and what you are going to get.

Nirav Merchant: And that's a great question. So, before -- the jet stream, when groups have to work together, as in the virtual organization and come together, they spent more time shipping data and coming to an agreement on standards, but this, all of it is already done for them, and if they come up with an idea and if we have to work using the same methods, it's all there. We don't have to try and decide, did I get the right method, and did they send me the right program. I am using exactly the same things, so that is a lot of reproducibility, accuracy, and the time of discovery is getting shorter, so you are bringing things to the market and research faster.

Ted Simons: Is there a learning curve dealing with all of this high techery. Sometimes learning how to streamline can be as big of a problem as streamlining itself.

Edwin Skidmore: Actually, that was the kind of the, the impetus for the atmosphere, is that we wanted to make using the cloud easier, so, literally, with the click of a button, a scientist or an engineer can start up a system on the cloud. So, we want to remove those barriers to accessing the cloud.

Ted Simons: So what kind of response are you getting for those in all sciences?

Edwin Skidmore: The demand has been tremendous. Actually, that was part of the reason why we were happy to be a part of jet stream.

Ted Simons: And talk about how jet stream now is going to expand? How many folks will be involved? How far can this go?

Nirav Merchant: The folks involved are collaborators in the advanced computer center, and we also have a very nice group of collaborators who are based serving areas where they have no access to high scale computing and they will be our partners and we'll be giving them the same infrastructure and access, and what we are hoping is what's called the federation, where other institutions, much like asu and others, will have their own cloud, and we will be able to connect them to form a better eco-system of academic cloud systems.

Ted Simons: A lot of activity going up in the cloud. Any security concerns with all of this?

Nirav Merchant: Absolutely. I think that the security is a concern for anything that is online now. We are very conscious of that, and one of our collaborators in this, is going to be focusing on how to manage the identity access, so we spend a lot of time trying to make sure that the systems are protected in ways that the Cyber security for it is not going to compromise anything that is in there. So, you know, we keep an eye out for things. But security is a concern.

Ted Simons: Is that a battle to make sure that the security is in place?

Edwin Skidmore: Absolutely. With these new models, with new access, it's into these domains of clouds. There is always a challenge to make sure that we're as secure as possible, as well as making it as available to scientists.

Ted Simons: And quickly, another challenge, I would think, would be just keeping up with computer advances. Correct?

Edwin Skidmore: Correct. And we have a team at the University of Arizona that is fantastic, you know, who are developing atmosphere and propelling it forward, and they ensure that we keep up to date.

Ted Simons: And you are confident that you will be keeping up to date with all of this?

Nirav Merchant: Absolutely, and that's the purpose of cloud. You don't really care about what is below. You are focusing on what you want to do. As the computers get faster, you get better results. You don't worry about the new model below it.

Ted Simons: Make sure the cloud stays up there and doesn't come crashing down.

Nirav Merchant: Right.

Ted Simons: Very good. Thanks for joining us.

Edwin Skidmore: Thank you.

In this segment:

Nirav Merchant:Co-Principal Investigator and Director, iPlant Collaborative and Bio Computing at Arizona Research Laboratories; Edwin Skidmore:Assistant Director of Infrastructure and Director, iPlant Collaborative and "Atmosphere," Cloud Computing Platform at University of Arizona;