ViewNet

A Senior Thesis by Prerna Ramachandra ’14 (COS)

It is no overstatement to say that improving UI (the user interface) is on the minds of the top in tech today, and that includes Princeton senior Prerna Ramachandra. But Prerna isn’t just jumping on the bandwagon of making the user experience better through design and functionality — she’s helping to lead the movement, specifically when it comes to software-defined networking. This computer science concentrator’s senior thesis, “ViewNet: A Visualization Tool for Software Defined Networks,” will likely be the first open-source visualization tool for finding and debugging networks.

Prerna explains to me the change that is coming to networks through an analogy. Think about your computer and its CPU, she suggests. That is the type of relationship we are focused on with networks, but now we want to separate the control (the decision-making) pane (the switch and, in our analogy, the CPU) from the router or data pane. Traditional computer networks are defined just by hardware (both router and switch), but software-defined networking, as you might have guessed, wants to make control panes all software. Now the push is to standardize this and to make visualization tools able to interact with these new software-defined control panes.

But, I have to ask, what could the motivations be for such a move throughout the tech world? To this, Prerna says, there are many answers, and, to my surprise, they touch more than just UI and UX (user experience). With software-defined networking, we can separate the person responsible for the rules of a network from the person responsible for debugging it. We can eliminate hardware and software dependence and enter a system similar to the cloud. Although in the past one was previously limited to using the tools of a company (for example, CISCO) that sold a router, software-defined networking allows the user to choose. With related visualization tools like Prerna’s thesis, there also arises the possibility of improving the experience and effectiveness of network debuggers by allowing them to study visual data rather than lines and lines of command-line prompts.

In face of such an interesting, unique and forward-thinking research topic, I have to ask Prerna from where the inspiration to pursue such a thesis came. She tells me that her introduction to software-defined networks came while doing research at Princeton over the summer after her freshman year. However, such an early introduction to the field doesn’t mean Prerna always knew she would make this her thesis. In fact, Prerna originally planned to pursue a different area of research, but, after taking a class her senior fall in human computer interface technology (with Professor Jennifer Rexford), this passionate-about-design engineer knew creating a visualization tool for software-defined networking was what she wanted to do.

Although Prerna did not have to write a thesis to earn her B.S.E., she was always attracted to the idea of doing so and felt it to be an “integral part of the Princeton experience.” She recommends that other engineers at least consider doing so provided they both pick a topic that really motivates them and get started early. Through her conversations about her thesis with advisor Professor Brian Kernighan, her communications with Professor Rexford and just through her own independent research, Prerna says she has learned so much by pursuing her thesis. From studying web development and writing in Javascript for the first time to having to really examine design-related questions such as “What is intuitive to the user?” Prerna would certainly characterize the senior thesis process as one-of-a-kind and exciting, if also an exhausting and challenging experience. Prerna hopes “ViewNet” will help impact the standardization of software-defined networking visualization tools and change the way those it affects will be able to approach their work, and we can’t help but think it will.

Update: Prerna Ramachandra is now a Product Manager at Intuit in Mountain View, CA.