Platforms Enable the Future

Platforms are important because they are synergistic with applications: new platforms enable new classes of applications, and new classes of applications often require the creation of new platforms. Many of the greatest revolutions in computing over the last 50 years were triggered by the development of new platforms and their associated applications.

Example Platforms

& the Applications they Enabled

Relational databases enabled enterprise applications.
The World-Wide Web enabled search, e-commerce, and social networks.

MapReduce enabled BigData applications.
Graphical Processing Units (GPUs) enabled image generation, animation, and machine learning.

Platforms Come In a
Variety of Shapes & Sizes

For example, remote procedure call libraries such as gRPC are smaller platforms that make it easy to communicate between processes running on different machines. They have been essential in enabling large-scale datacenter applications.

Our Mission Is to Develop New Platforms for the Future

By doing this, we hope to enable new kinds of applications that simply aren’t feasible today. We work on a variety of platforms, large and small, hardware and software. Most of our work is in the areas of networking, systems software, and datacenters.

The Faculty of the Stanford Platform Lab Have a Long & Distinguished History of Developing Influential New Platforms

Here are a few examples of past work by the Stanford Platform Lab faculty. Although we cannot predict what new platforms will be coming out of the Platform Lab in the future, we believe they are likely to be as significant as the ones below:

Virtual Machines

Virtual machines provide a powerful mechanism for managing and sharing computing resources in datacenters. The modern notion of a virtual machine grew out of research by Stanford Platform Lab faculty, resulting in the founding of VMware and the creation of a new virtualization industry.

Virtual Machines

Software Defined Networking (SDN)

Software Defined Networking (SDN) has disaggregated networking infrastructure into a few programmable platforms that enable operators and third parties to innovate and customize their networks. The new platforms include OpenFlow, which makes the forwarding plane programmable, and P4, which enables programmable packet processing pipelines.

Tcl Scripting Language

The Tcl scripting language made it easy to create a variety of powerful command languages. When combined with the Tk toolkit, it lowered the complexity bar for creating graphical user interface applications.

Spark

Spark provides a new paradigm for creating large scale “Big Data” applications in datacenters; it is now widely deployed and used throughout the industry.

Gg

Gg is a new framework that helps people execute everyday applications using thousands of parallel threads on “serverless” platforms to achieve near-interactive completion times.

Raft

Raft provides a new mechanism for achieving consensus among a cluster of servers. It has simplified the development of highly reliable distributed applications.

TinyOS

TinyOS is an open source operating system with an extremely small form factor intended for low-power devices; it has simplified the development of applications for devices that would previously have been very difficult to program.

NetFPGA

NetFPGA is a programmable platform for line rate networking research, education and real world applications. There have been multiple versions of it for 1G, 10G and 100G networking.

How We Work

The Stanford Platform Lab consists of about 10 Stanford faculty working in a variety of areas in Computer Science and Electrical Engineering, along with dozens of PhD students and other researchers. At any given time we typically have a few overall themes that cover much of our active work, along with a variety of smaller projects.

All of our work is open.
We enjoy collaboration, especially with industrial partners.
We don’t just do interesting research; we also build practical systems that can impact the actual practice of computing and networking.