🚀  CDAC @ IRUS 

Title : High-throughput Network Traffic Analytics Platform

When : December 21, 2024, 10:30 AM – 3:00 PM at Robusta

Organizer: Shivakumar Sastry, Senior Advisor, C-DAC Bangalore

Steering & Coordination: Dr. S D Sudarsan, Executive Director, C-DAC Bangalore

Technical Contributions:  Dr. Balaji Rajendran, Dr. Muraleedharan, Ms. Deepika H.V, C-DAC, Bangalore.

C-DAC-Bangalore aims to launch a unique mission-mode platform-project to deliver a high throughput network traffic analyzer and develop a sustainable national ecosystem to maintain and upgrade this family of products into the foreseeable future. The platform is an integrated hardware, software and firmware framework for a family of network traffic analyzers that are highly customizable to address the needs of multiple verticals; the platform will be continuously upgraded to remain abreast of the underling computing and networking technologies into the foreseeable future. The platform cannot be designed and evolved without a national ecosystem for talent and resources. Thus, it is a primary objective of this mission-mode activity to create a sustainable ecosystem that includes talent for architecture, design, tools, processes, requirements, verification and validation, manufacturing, operations and maintenance. A particular focus will be in areas such as high-speed digital design, processor design, ASIC design and high-performance software systems.

Perspectives: A large mission-mode activity such as the one above admits multiple perspectives. From the C-DAC perspective, it is an outcome-focused, applied research, activity that will address a critical national need and generate protected intellectual property (IP) to eliminate national dependence on a weak supply chain. From the perspective of the nation, this activity is a local talent creation engine.

For the partner industries – there are two perspectives: a) for established companies, it is a research commercialization activity, and b) for startups it is a new opportunity to partner with C-DAC to create shared IP. Finally, from the perspective of partner academic institutions, it is an activity that can provide its faculty and staff an opportunity for hands-on engineering that is rigorous, well-managed and forward looking – to clarify further: the area of network processors is already well-researched, the opportunity to publish incremental academic ideas is extremely limited and will be strongly discouraged.

The Need: The cumulative downlink capacity for India is around 1500 TB. There is an urgent need to secure this network traffic against both malicious and benign actors. Currently, C-DAC has a software-only product for network traffic analysis that has been indigenously developed. Deep packet inspection is a core feature of this product, this platform is likely to deliver a throughput of about 25 GBPS with organic advances in processor architecture. In the US, commercial products deliver throughput between 400 and 800 GBPS and in September 2023, a 1.2 TBPS prototype was demonstrated in China. Clearly, there is a need to design and engineer a high-throughput platform that can support the digital aspirations of the nation while assuring strong network security that cannot be compromised by weaknesses in the supply chain.

Challenges: The primary national challenge is to urgently secure current and future Indian networked systems against cyberattacks with indigenous intellectual property. This activity is challenging along many dimensions:

1. The scaling of throughput from about 25 GBPS to the TB levels can only be accomplished by integrated teams that involve FW, HW, SW, NW engineers and designers; currently, there is no integrated mechanism to identify, assemble and manage such a diverse team.
2. Given that the total downlink capacity is going to grow for India, and India’s data must be secured by indigenous IP, the market size in India for TB range network traffic analyzers is moderate. Further, India-based companies face several hurdles to penetrate the global computing and networking markets. For these reasons, it is unlikely for a private industry to pursue such a project even when it addresses a critical national need. Thus, C-DAC would like to step up and lead this effort both to secure the nation’s digital aspirations and to revitalize a talent pool that can contribute to national advances beyond this challenge.
3. The technical aspects are formidable. Current network traffic analyzers work offline or in near real- time modes. Scaling this to a real-time or near real-time operation requires new approaches and judicious integration of emerging AI approaches. A traffic analytics platform with about 2 TBPS throughput is recognized as a global challenge for 2025. Thus, it is important to assemble a highly talented team to conceive, plan and deliver such an ambitious program.
4. The available talent pool in India for several critical areas including advanced networking, advanced processor design, ASIC, and advanced software is limited. Thus, it is critical to groom a talent pool in these areas in India. In recognition of competitive marketplace and the nature of the platform-project, critical members of the team must be compensated on par with the industry standards.

This discussion emphasizes that the identification and retention of a talented team is the primary requirement for planning and executing the proposed mission-mode activity. For this reason, we will engage potential industry, academic, research labs, startups and other community partners through the IRUS Symposium in December 2024. Action items from this meeting, and other such planned partner- engagement activities in March and Summer 2025, will enable C-DAC to identify and assemble an effective and talented team for this critical platform-project. Specific goals for IRUS@HiPC, December 2024 are:

1. Socialize the desired outcomes to potential industry, academic, research labs and other community partners.
2. Identify key national leaders who can lead proposal preparation and submission activities.
3. Establish a shared vision for industry stakeholder (potential future customer) needs.
4. Align partner needs and goals with program outcomes.
5. Plan next partner-engagement activities.

IRUS@HiPC’24 Tentative Agenda:

1. Inauguration
2. Overview of the HT Network Traffic Initiative
3. India Partner Network
4. Lightning Talks
5. Panel Discussion
6. Wrap up and Next Steps