To cost-effectively deploy ANI, ESnet partnered with Internet2—a consortium that provides high-performance network connections to universities across America—which also received a stimulus grant from the Department of Commerce’s Broadband Technologies Opportunities Program.
So far more than 25 groups have taken advantage of ESnet’s wide-area testbed, which is open to researchers from government agencies and private industry to test new, potentially disruptive technologies without interfering with production science network traffic. The testbed currently connects three unclassified DOE supercomputing facilities: the National Energy Research Scientific Computing Center (NERSC) in Oakland, Calif., the Argonne Leadership Computing Facility (ALCF) in Argonne, Ill., and the Oak Ridge Leadership Computing Facility (OLCF) in Oak Ridge, Tenn.
“No other networking organization has a 100-gigabit network testbed that is available to researchers in this way,” says Brian Tierney, who heads ESnet’s Advanced Networking Technologies Group. “Our 100G testbed has been about 80 percent booked since it became available in January, which just goes to show that there are a lot of researchers hungry for a resource like this.”
To ensure that researchers will use future 100-gigabit effectively, another ARRA-funded project called Climate 100 brought together middleware and network engineers to develop tools and techniques for moving unprecedentedly massive amounts of climate data.
Approximately 13.7 billion years ago, the Universe was almost homogenous — meaning that every location in the cosmos was similar. Today, this is no longer the case. This simulation starts from a nearly homogeneous Universe and shows how the it has changed over billions of years. Performed on 4,096 cores of NERSC’s “Hopper” system with the Nyx code, this movie was generated with over 5 terabytes of data and was transferred to the SC11 Conference exhibit floor in Portland, Ore., last November, over ESnet. The video on the left shows the simulation streaming on a 10 Gbps link, while the one on the right shows the same model streaming on a 100 Gbps link. These simulations were generated by Prabhat (LBNL).
In 2024, the most powerful radio telescope ever constructed will go online. Comprising 3,000 satellite dishes spread over 250 acres, this instrument will generate more data in a single day than the entire Internet carries today. Optical fibers will connect each of these 15-meter-wide (50 ft.) satellite dishes to a central high performance computing system, which will combine all of the signals to create a detailed “big picture.”
“Given the immense sensor payload, optical fiber interconnects are critical both at the central site and from remote stations to a single correlation facility,” says William Ivancic, a senior research engineer at NASA’s Glenn Research Center. “Future radio astronomy networks need to incorporate next generation network technologies like 100 Gbps long-range Ethernet links, or better, into their designs.”
In anticipation of these future networks, Ivancic and his colleagues are utilizing a popular high-speed transfer protocol, called Saratoga, to effectively carry data over 100-gigabit long-range Ethernet links. But because it was cost-prohibitive to upgrade their local network with 100-gigabit hardware, the team could not determine how their software would perform in a real-world scenario—that is, until they got access to the ANI testbed.
“Quite frankly, we would not be doing these speed tests without the ANI testbed,” says David Stewart, an engineer at Verizon Federal Systems and Ivancic’s colleague. “We are currently in the development and debugging phase, and have several implementations of our code. With the ANI testbed, we were able to optimize and scale our basic PERL implementation to far higher speeds than our NASA testbed.”
Meanwhile, Dantong Yu, who leads the Computer Science Group at Brookhaven National Laboratory, used the ANI testbed to design an ultra-high-speed, end-to-end file transfer protocol tool to move science data at 100 gigabits per second across a national network.
“A network like ANI may be able to move data at 100 Gbps, but at each end of that connection there is a host server that either uploads or downloads data from the network,” says Yu. “While the host servers may be capable of feeding data into the network and downloading it at 100 Gbps, the current software running on these systems is a bottleneck.”
According to Yu, the bottlenecks are primarily caused by the number of times the current software forces the computer to make copies of the data before uploading it to the network.
“Initially I was testing this protocol at a very local lab level. In this scenario transfers happen in a split-second, which is far from reality,” says Yu. “ANI allowed me to see how long it really takes to move data across the country, from East-to West Coast, with my software, which in turn helped me optimize the code.”
The Next Steps
Within the next few months, the official ANI project will be coming to an end, but the community will continue to benefit for decades to come from its investments. The 100-gigabit prototype network will be converted into ESnet’s fifth-generation production infrastructure, one that will be scale to 44 times its current capacity. ESnet will also seek new sources of funding for the 100-gigabit testbed to ensure that it will be available to network researchers on a sustained basis.
“Since its inception, ESnet has delivered the advanced capabilities required by DOE science. Many of these capabilities are cost-prohibitive, or simply unavailable, on the commercial market,” says Bell. “Because our network is optimized for the needs of DOE science, we’re always looking for efficient ways to manage our large science flows. ESnet’s new 100-Gigabit network will allow us to do that more flexibly and morecost-effectively than ever.”
If you liked this article, please give it a quick review on ycombinator or StumbleUpon. Thanks