Are you passionate about learning and open minded about the way that networks are built? Do you have a passion for organizing and visualizing data to aid in the understanding and development of network systems? Consider the Research and Development team of Berkeley Lab’s Scientific Networking Division. At the core of the Division is ESnet - the Energy Sciences Network.
ESnet’s mission is to accelerate science by delivering unparalleled networking capabilities, tools, and innovations. ESnet interconnects the US national laboratory system, is widely regarded as a technical pioneer, and is currently the fastest science network in the world. We are a dynamic organization, highly motivated and focused on results. We are working at the leading edge of software-defined networking, network knowledge plane, dynamic network infrastructure, network visualization, network knowledge plane, multi-domain and multi-layer architectures, deep learning etc. The successful student will be the one that brings strong and diverse coding skills and is very self motivated.
Project One
Title: Experimenting with Deep Reinforcement learning and parallel computing
Term: Spring/Summer 2020
Project Mentor: Mariam Kiran
Required skills: Experience with building OpenAI gym and DQN algorithms
Abstract:
Facilities and instruments are exploring methods to develop reinforcement learning approaches that can be deployed to allow instruments to learn themselves. In this work, we will be exploring methods on how we can develop and test DQN algorithms with Gym models created to mimic the facility. Additionally, these models will be deployed on parallel architectures to encourage quick processing and stability reach.
This project will look at understanding how reinforcement learning approaches can be deployed with real systems and the challenges faced in running the code.
Project Two
Title: Network Telemetry Processing at 100Gbit/s and Beyond
Term: Summer 2020
Project Mentor: Richard Cziva
Required skills: Networking and Linux fundamentals, experience of one or more of today’s packet processing technologies, such as: DPDK, XDP, eBPF, P4 (preferred) and similar techniques. Experience with Barefoot Tofino or Netronome platforms is an advantage. Programming experience in Go is a plus. *PhD student in networked systems is preferred*
Abstract:
The next generation of ESnet’s network (ESnet6) will introduce `High-Touch Services`, an internal hardware and software solution to deliver enhanced, high-speed network services for network operators and users in real-time. These high-touch services will enhance user experience, collect service quality metrics (e.g., TCP performance monitoring), monitor network security (e.g., delay and flow monitoring) and enable real-time network debugging and packet analysis in the world’s fastest science network.
You will be working with ESnet’s high-touch team, designing, implementing and further enhancing already implemented high-touch services and associated management software. You will have the option to explore state of the art programmable hardware platforms for high-speed packet processing, such as Barefoot Tofino 6.5TB switches and Netronome cards. This project gives an opportunity to learn and use P4, a programming language designed to allow programming of packet forwarding planes.
If you are someone who is excited about high-speed packet processing and network programmability on a multi-100Gbps fiber optic backbone network that stretches across the country and beyond - this internship is for you!
Project Three
Title: Failure Modeling of International Science Networks
Project Mentor: Richard Cziva
Term: Summer 2020
Required skills: Strong networking knowledge, experience in data analysis and scripting. *PhD student in networked systems is preferred*
Abstract:
International science networks, such as ESnet, move hundreds of Petabytes of data a month between continents, interconnecting instruments with users, storage and processing facilities. While the best quality of service is targeted, sometimes things do go as planned and network failures occur.
This project will be looking at understanding where these failures could occur in similar networks and what impact they have. We will investigate years of operational data collected at ESnet’s various systems and apply natural language processing on unstructured data such as operational emails. With the information derived, we will create a ESnet specific taxonomy of network failures and investigate if we can see correlations between events logged in different layers of the network. Furthermore, we will try to build a prediction for some of the failures.
Project Four
Title: Optical quality of transmission modeling of ESnet5 wavelengths using GNpy
Project Mentor: Chris Tracy
Term: Summer 2020
Required skills: python, DWDM, strong engineering background
Abstract:
GNpy (https://gnpy.readthedocs.io/en/master/) is a new open-source tool released by the Telecom Infra Project (TIP) which estimates the quality of transmission performance of optical signals over dark fiber networks using a Gaussian Noise model to determine the nonlinear impairments. We want to use GNpy to model an existing ESnet5 production optical span, and compare the performance estimation of the open-source Gaussian Noise model to closed-source proprietary vendor tools.
You will use real operational data from one of the fastest science networks on the planet, and work with bleeding edge open-source tools in the area of open optical networking. You will also have a chance to see how modern large-scale optical networks are designed and operated, and how this field is rapidly evolving.
Project Five
Title: Network Telemetry Collection and Analysis Systems
Project Mentor: Sowmya Balasubramanian, Andy Lake
Term: Summer 2020
Required skills: Strong CS Fundamentals - Data structures, algorithms, fluency in at least one language - Python, Java or Go, Familiarity with relational and/or non-relational databases. Other Desired Skills: Previous experience with web services development preferred, Some UI knowledge (Experience with React is a plus).
Abstract:
ESnet’s Measurement and Analysis Team in Software Engineering is looking for an exceptional software intern to work on various parts of our time-series network measurement data collection and analysis systems. This telemetry data provides visibility into the health of the network in near real-time and helps to troubleshoot any network issues. Work may involve designing parts of collection system, implementation of its APIs and/or visualization libraries.
Project Six
Title: Developing Predictive Analytical Models and Methods (using Clustering and Extrapolation) for Proactive Network Health Management
Term: Summer 2020
Project Mentor: Cody Rotermund
Required skills: Networking and Linux fundamentals along with Python and general shell scripting familiarity. Experience with machine learning and predictive algorithms are an advantage.
Abstract:
ESnet gathers an impressive amount of data in the day to day operation of the ESnet network, including hardware health/error/event logs, bandwidth utilization statistics, real-time network equipment telemetry, and network monitoring alarms. Moving forward into the next iteration of ESnet’s production network, there will be even more data captured from sources such as our optical networking platform as well as our network orchestration and automation solutions.
ESnet believes that this data can not only be used to tell us about our current state of the network and it’s past performance, but help us to predict hardware and software failures in the future. With reasonable forewarning, in this project, we want to explore if ESnet would be better able to perform planned maintenance and remediation on our network systems to minimize overall downtime and move ESnet from a reactive network management model to a more proactive one.
You will be working with ESnet’s Network Engineering and Machine Learning teams, to first identify and label the sources of historical and real-time data; then working with engineers to correlate this data to known outages and unplanned disruptive network events. Finally, using this well-labeled data and correlated impacts to the network, you will work to develop predictive models that will help to forecast outages and network performance issues. Ultimately, the goal will be for ESnet to be able to take action in advance of highly likely network disruptions and reduce unplanned outages based on the predictions that you have helped to develop.
As part of the application process, please submit a statement explaining which project you prefer to apply to.
Notes:
The Spring 2020 Term is 16 weeks (1/6/2020 - 4/24/2020). The Summer 2020 Term is 12 weeks (6/1/2020 - 8/21/2020). Student participation requires 40 hours per week commitment for Summer appointment, and 20 hours per week commitment for the Spring appointment. A "late start" date can be considered for academic reasons.
The student assistant appointment can be renewed based upon satisfactory job performance, continuing availability of funds and ongoing operational needs.
There are multiple openings for this position.
Salary will be predetermined based on student step rates.
This position may be subject to a background check. Any convictions will be evaluated to determine if they directly relate to the responsibilities and requirements of the position. Having a conviction history will not automatically disqualify an applicant from being considered for employment.
Work will be primarily performed at Lawrence Berkeley National Lab in Berkeley, CA or the Champaign, Illinois office.
Learn About Us:
Working at Berkeley Lab has many rewards including a competitive compensation program, excellent health and welfare programs, a retirement program that is second to none, and outstanding development opportunities. To view information about the many rewards that are offered at Berkeley Lab- Click Here.
Berkeley Lab (LBNL) addresses the world’s most urgent scientific challenges by advancing sustainable energy, protecting human health, creating new materials, and revealing the origin and fate of the universe. Founded in 1931, Berkeley Lab’s scientific expertise has been recognized with 13 Nobel prizes. The University of California manages Berkeley Lab for the U.S. Department of Energy’s Office of Science.
Equal Employment Opportunity: Berkeley Lab is an Equal Opportunity/Affirmative Action Employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, age, or protected veteran status. Berkeley Lab is in compliance with the Pay Transparency Nondiscrimination Provision under 41 CFR 60-1.4. Click here to view the poster and supplement: "Equal Employment Opportunity is the Law."
