The new Student Activities Funding Engine, also known as SAFE, went live in November 2013. SAFE is a custom-built solution developed by OIT in close partnership with the Office of the Provost, the Office of the Vice President for Campus Life, and the Office of the Dean of the College.
The new SAFE system coordinates the process for funding undergraduate student activities. Undergraduate students use the SAFE portal to search and apply for University funding. Funding sources, such as academic departments, administrative offices and campus life groups use the portal to post funding opportunities and to manage the funding process from application review to payment. SAFE brings both of these groups together and brokers student funding needs with funding opportunities, and helps ensure fair and equitable distribution of funds to the greatest number of students.
For students, SAFE provides a clearinghouse to the various funds that are available to support undergraduate student activities off campus. This gives all students access to the same information providing equal opportunities to apply for financial resources. The types of activities currently coordinated in SAFE include senior thesis, summer study aboard, internships and fellowships, and independent projects and other research. Several departments already observed an increase in the number of students applying for available funding. Humanistic Studies experienced nearly 40% more applications this year than in each of the last four years and attributes SAFE for reaching a broader eligible audience.
For the administrators of funding resources and for offices with funding opportunities, SAFE increases transparency and improves the efficiency of the funding process. Through SAFE, funding offices can interact with other possible funders to ensure that projects are fully funded whenever possible, and that available funding is awarded most efficiently and supports the greatest number of students. The coordination between funding offices can also help eliminate the possibility of over-funding projects.
Before SAFE, no method existed for reporting on cross-campus funding allocations. SAFE enables the University to track and report on how funds are spent, including endowed gift funds at the University on behalf of students. SAFE also offers efficiencies: funding opportunities are entered once and republished at the frequency of the award (semester, annually, other); forms are auto-populated with known data; the self-help portal provides comprehensive information about regulations, procedures and resources for undergraduate student funding, streamlining the interaction between the student and funding offices (forms are readily available online) and moving requests more quickly from student application to award.
Currently, there are 198 funding opportunities in the SAFE system. Students submitted 1,142 applications requesting funding. To date, 970 student projects received funding totaling $2,986,257. Of the 172 projects that did not receive funding, many were secondary or duplicate requests submitted by students who received funding for another project or from another funder. The students who used SAFE to request funding represent more than 35 concentrations of study.
Plans are in progress to extend SAFE for use in supporting additional processes, including funding for on-campus activities and funding for group activities that are on- and off-campus. The system is also being developed further to include points of integration with Concur, the University’s new online travel booking and expense management system.
The Princeton Institute for Computational Science and Engineering (PICSciE) has expanded once again, reaching a new level capable of sustaining world-class computation-based research. Recent additions and enhancements located in our state-of-the-art High Performance Computing Research Center (HPCRC) on the Forrestal Campus ensure that high-performance computing resources at Princeton continue to meet the expanding and specialized computing needs of a growing user base of Princeton faculty, researchers, and students.
Tiger - Princeton’s largest computing system
Princeton’s newest high-performance computing cluster, Tiger, is by far the University’s largest research computing system to date. Tiger increases Princeton’s overall computational capacity by a factor of 3 to 4. It is also likely that the overall performance of the Tiger cluster, once both phases of the cluster are in production, would rank the supercomputer within the top 200 on the TOP500 Supercomputer Sites list, should Princeton choose to run and submit a benchmark for the system.
The first phase of the cluster, now in production, is composed of 592 servers or “compute nodes,” each having 16 computing cores and 64 to 128 gigabytes of RAM. The second phase will add 52 nodes to the cluster, of which 50 nodes each will contain four NVIDIA Kepler K20 graphical processing units (GPU) and the remaining two will contain four Intel Xeon Phi processors. Tiger accelerators will greatly enhance the performance of certain scientific and engineering models, in some cases by more than a factor of ten. The second phase will be in production at the end of March 2013.
The Tiger cluster was funded with central funds, departmental funds, funds from individual faculty, and a National Science Foundation (NSF) Computer & Information Science & Engineering (CISE) Computing Research Infrastructure (CRI) grant, A Platform for Data-Parallel GPU Computing at Princeton, with David August as the Principal Investigator.
Disk capacity - now 10 times greater
In October of 2012, Princeton was awarded an NSF Major Research Instrumentation (MRI) grant, Acquisition of a Shared Parallel High Performance Storage System to Enable Computational Science and Engineering, with Curt Hillegas as the Principal Investigator. Funds from this grant and required co-investment funds were used to purchase a storage system that increases disk capacity and i/o performance by a factor of more than 10. The system provides 2.6 petabytes of formatted storage and aggregate i/o bandwidth exceeding 12 gigabytes per second. This system will be in full production by the end of March 2013.
Orbital – a system for distributed computing over many nodes
The computational cluster Orbital was purchased and installed in January of 2012. This system, which was funded primarily by the Astrophysics department (Anatoly Spitkovsky) and the Physics department (Frans Pretorius) is comprised of 308 compute nodes, each with 12 compute cores and 48 gigabytes of RAM. This cluster is distinguished by its powerful Infiniband network that is configured to allow any two nodes to communicate at full speed with each other regardless of the communication between the other nodes in the cluster. This configuration is particularly tuned to science and engineering models that need to be spread across many compute nodes. At the time of its purchase, Orbital was the most powerful cluster in Princeton’s suite of high-performance systems.
Hecate – one of the largest shared memory systems in the world
Princeton was awarded an NSF MRI grant, Acquisition of a Shared Memory High Performance Computer for Modeling and Data Analysis in the Mathematical and Physical Sciences, with Roberto Car as the Principal Investigator. Funds from this grant and required co-investment funds were used to purchase and install Hecate, a SGI UV1000 system. Hecate was the first system installed in the new High Performance Computing Research Center (HPCRC) on the Forrestal campus in January 2012. This specialized system is one of the largest shared-memory systems in the world, with 1536 processing cores and 12.1 terabytes of RAM. The system runs a single operating system and any application running on the system can use all of the memory in the system. Hecate provides significant performance advantages for science and engineering models that require tight coupling between the processors.
The strong collaboration between OIT and PICSciE to manage Princeton’s central high-performance computing systems and to support the researchers who use them continues to expand and grow. This collaborative approach has extended to include system administration and support participation from academic departments. Since 2009, the (Scientific Computing|Super Computing) Administrators Meeting (SCAM) group has been bringing the central group together with academic department system administrators who also support this type of research, which helps foster a number of relationships leveraging the special expertise of departmental computing support staff to improve Princeton’s central high-performance computing environment.
This past August, SuperStorm Sandy demonstrated the success of recent IT infrastructure improvements made to ensure the high reliability and continuous operation of IT services at Princeton. Central IT services, research computing, and departmental IT services housed in the University’s central data center facilities operated without interruption or incident despite the 30-hour long power outage caused by the storm.
Before improvements began in 2008, the University’s IT infrastructure depended on a single data center. Power backup mechanisms at this site were designed to provide backup power for servers and storage systems for up to 15 minutes and run the central network for up to 90 minutes. As more services came to rely on the University’s central network, the need to provide service redundancy and extended power backup became more pressing.
With the conversion of New South to a fully functioning data center facility in 2009 and the recent opening of the state-of-the-art High Performance Computing Center (HPCRC) in 2012, central servers and data storage are now located in buildings equipped with onsite, backup power that protects these services from extended utility power outages. Generators at both locations start up without manual intervention in the event of a power failure to ensure continuous service.
When power was lost during the recent storm, the HPCRC switched automatically and seamlessly to its backup generators. The data center facility in New South continued functioning with power provided by the University’s co-generation system. A risk that remained was the lack of backup power for the University’s network core at 87 Prospect. A temporary generator was deployed before the storm to support the network core throughout the power outage. A project to install a permanent generator at this site is now underway and is expected to complete in early 2013.
The continuous operation of the University’s IT systems and services during Superstorm Sandy supported the University in its storm recovery efforts in many ways. In particular, it provided continuity in the University’s communication systems and the ability for Facilities to monitor the power status of buildings on campus throughout the storm. Princeton’s web presence and email communications were uninterrupted, which allowed important communication to all campus constituents and among the staff charged with emergency operations. Investments in the University’s IT infrastructure that make it less vulnerable to utility power interruptions ensure that research, administrative, and communications systems remain continuously available to support University operations.
The Princeton University network has become an indispensable tool for the basic operation of the University. As such, the enhancement of Princeton’s network and telecommunications services continues to be a top priority in OIT.
Wireless network upgrade
Over the past five years, the number of unique devices connecting to Princeton’s wireless network each year has grown from 42,000 devices to nearly 141,000, representing a 238% increase. From FY08 to FY11, the number of connections increased an average of 29% per year. Last year, the wireless network experienced the greatest increase, with 55% more connections than the previous year.
To help ensure the reliability and future growth of the Princeton wireless network, OIT is working on several wireless network projects that increase network capacity and support a greater number of simultaneously connected devices, at faster speeds.
Access point upgrade
OIT is currently engaged in a multi-year project to upgrade the University’s wireless network by replacing the devices (access points) that provide wireless coverage. The purpose of this upgrade is to provide the higher bandwidth available through newer network technology (802.11a/b/g/n) and to improve our ability to control the level and quality of service. OIT has partnered with Aruba Networks on this project. As part of this project, each existing access point is being replaced with multiple access points, increasing access speeds and the network capacity by almost an order of magnitude. To date, the project is 40% complete, with just over 20 administrative and academic buildings and 38 undergraduate and graduate dorm buildings fitted with the upgraded wireless service.
In the coming year, OIT is scheduled to complete work on a number of administrative and academic buildings and, in the summer of FY13, will upgrade all remaining undergraduate and graduate dorms. The campus-wide upgrade to the wireless network is expected to complete in FY16.
Network Address Translation (NAT) installation
The multitude of devices connecting to Princeton’s wireless network coupled with a limit of 16,000 campus internet addresses (IP) available for wireless service put a significant strain on the ability to provide network access to these devices. Once a supplemental service, the wireless network has become a primary network service on campus, supporting more than 40,000 registered wireless devices. It is not uncommon for students, faculty and staff to each have two, three, or four devices that rely on Princeton’s wireless infrastructure.
To support this recent proliferation and the anticipated exponential growth of devices connecting to Princeton’s wireless network, OIT deployed Network Address Translation (NAT) devices. This upgrade increased the number of wireless devices the network is able to concurrently serve from 16,000 to over 100,000 devices.
Increased Internet bandwidth
The University has two Internet Service Providers (ISPs) that together help balance network traffic and ensure service continuation should one of the services fail. During the past academic year, usage of the University's two Internet links regularly exceeded 50% of their peak capacity. At this level of traffic, if one of the connections failed, the other alone would not be sufficient to meet the Internet access needs of the University.
In FY13, OIT is working to double the speed of the commodity Internet links to increase service capacity and ensure service continuation in the event one of the connections should be unavailable.
OIT has prepared the campus network infrastructure for participation in the "eduroam" wireless access service. Eduroam is an initiative of Universities from around the world that is designed to facilitate collaboration by providing a simple way for faculty, students, and staff to get secure access to the Internet while visiting participating institutions. Eduroam provides a benefit for members of the Princeton community as they travel the world; it also provides scholars and colleagues from participating institutions secure and convenient access to the wireless network while visiting Princeton. The connectivity is instantaneous and only requires login and password from the user.
The availability of eduroam has added benefits. Eduroam simplifies network administration by eliminating the need to provision visitors from other institutions on the Princeton network. It also further supports our security initiative by providing an encrypted wireless connection.
Cellular infrastructure improvements
OIT continues to work with cellular providers to install cell towers on building rooftops to improve cellular reception around campus and the surrounding community. The campus now has eight towers that provide service for AT&T, Verizon, T-Mobile, Sprint or Cricket. An on-campus Metro/PCS cell tower is in the works.
OIT also continues to build-out a campus Distributed Antenna System (DAS) to provide better cellular reception inside buildings.
The Office of Information Technology (OIT) continues to investigate the use of outsourced IT services for the campus community. Recent changes in student e-mail and IT infrastructure services illustrate OIT’s commitment to offering alternative sourcing services, when they best match the needs of the University in a cost-effective manner.
Student E-mail Services
In February 2012, OIT partnered with the Undergraduate Student Government (USG) to pilot Google Apps for Education and Microsoft Office 365 e-mail and collaboration services. At the conclusion of the pilot, the USG issued a recommendation to adopt Google Apps for Education as the default offering for undergraduate student e-mail and collaboration tools and to also make available Office 365 for students who prefer a Microsoft-based solution. The University endorsed these recommendations.
OIT began transitioning to the cloud-based service in May 2012 and created “Google Apps at Princeton” accounts for the incoming class of 2016. By early this fall, the e-mail accounts of all undergraduate students will be moved to the new service. Princeton is also supporting a Microsoft solution for e-mail and calendaring as an alternate offering. These new services offer significant enhancements to students, including sizeable e-mail and storage quotas and strong support of mobile devices. They also allow OIT to refocus resources to support new technology initiatives for students and faculty.
In FY13, OIT will work with representative faculty, graduate student and administrative staff to pilot these same services for possible use by the rest of the institution.
IT Infrastructure Services
In September 2011, OIT partnered with Dell to supplement the level of security monitoring and enhance the protections for those University business systems that store confidential data. As part of this service, Dell Secureworks regularly provides reports and analyses of the layers of IT security around these systems, which allows OIT security staff to test and implement timely and stronger measures of protection.
In July 2012, OIT and Dell launched a joint project to evaluate Dell’s cloud-based service for enterprise servers and storage. The goal of the evaluation project is to test the security and network-speed requirements for using cloud-based servers and storage in the Princeton environment.
OIT is working closely with the McGraw Center for Teaching and Learning to record Princeton’s first set of online, interactive lecture offerings. The lectures are being offered through Princeton’s Coursera site at www.coursera.org/princeton. In FY13, nine online courses will be offered.
This past spring, Princeton offered its first online course through Coursera, Sociology 101, taught by Professor Mitchell Duneier. Nearly 40,000 participants enrolled in the course, sharing more than 12,500 posts in 1,800 discussion threads. Throughout the course, more than 25,000 unique participants viewed 235,900 video lectures. Sociology 101 was the first social science course offered through Coursera. 65 percent of the students enrolled in the course resided outside of the United States, with high representation from the countries of Brazil and India.
Two fall courses are in progress, Statistics One, given by Professor Andrew Conway, and Algorithms, Part I, given by Professor Robert Sedgewick. The algorithms course began with an enrollment of 65,600 students and is expected to grow. Other courses being offered in the fall include:
• A History of the World since 1300, with Professor Jeremy Adelman
• Computer Architecture, with Professor David Wentzlaff
• Networks: Friends, Money, and Bytes, with Professor Mung Chiang
Courses planned for the spring include:
• Algorithms, Part II, with Professors Robert Sedgewick and Kevin Wayne
• Analytic Combinations, Parts I and II, with Professor Robert Sedgewick
• Introduction to Sociology, with Professor Mitchell Duneier
In addition to the Cousera online offerings, this past spring OIT worked with the Council on Science and Technology to create an in-classroom recording of Neuro Science 101, with Professor Samuel Wang. Eight additional recordings of introductory science courses are planned for the fall. They will be used as teaching resources at Princeton and, as appropriate, beyond our campus.