In 
During day 2 of the felicitation of the ITNext 100 CIO winners event at Agra, Dr. Pramod Varma, Ex-Chief Architect Aadhaar & India Stack, CTO EkStep Foundation, started the day with his keynote address on the ‘Future of Infrastructure.’ He was asked a few pertinent questions, and Dr. Varma’s responses were significantly insightful. These are-
How will the cloud evolve in the near future?
All technology solutions are and will be available in the form of services. Whether it is an application, database, infrastructure, IoT, Machine Learning, Computer Vision, Natural Language Processing, Contact Center, and many more, cloud providers have commoditized or democratized these services with just a “tap,” Artificial Intelligence will be very prevalent in the form of cloud services.
What is your opinion on Infrastructure as a Code?
Multiple big enterprises/organizations struggle to make infrastructure (e.g., router, firewall, group policy) changes within their ecosystem. These changes are usually messy and error-prone since the dependencies are not clearly documented, and infrastructure becomes sophisticated yearly. The processes followed are also complex and time-consuming.
What is the future of Data Centers?
The majority of large & medium scale enterprises have migrated or started to migrate their data centers to the cloud. Hence, we will have the following characteristics-
a. Compute efficiently
b. Energy efficient
c. Heterogeneous cloud approach
d. Open Standards
e. Software-driven layers
Role of quantum computing incorporate
Dr. Varma mentioned that quantum computing will have a significant role in AI / ML-based solutions and platforms.
Energy efficiency must be considered when any organization thinks about leveraging quantum computing. Hence, when to start leveraging quantum computing is very important. Adopting a quantum computing ecosystem requires organizations to be fully green already.
It also requires quantum computers to accept open standards since organizations have their preferred set of products and partners.
What is the impact of Generative AI, LLMs (Large Language Models), and ChatGPT?
The majority of our knowledge resides in people’s brains or in text across portals or in PDF documents. The other problem is that this knowledge set is increasing every day, but no one is able to use it (leave about using it efficiently).
Translation of this knowledge to answer a specific query from any domain or industry is very time-consuming. Getting the correct answer for a specific question may take more than a day.
Dr. Varma gave an example of a farmer asking a niche question about agriculture in his language in their native language to a mobile device through an interface (or application). The device then converts voice to text; in turn, the text gets passed to LLM as a prompt, where LLM receives the right answer. The answer is then converted into voice, which is read back to the farmer through the same interface or application. Hence, knowledge will be democratized.
How can we ensure data security and privacy while maintaining trust in the system?
Dr. Varma gave a very intelligent and practical approach to adopting security and privacy. He clearly mentioned to bake (incorporate) all policies within the code and infrastructure.
We usually have variances between our policies and practices. However, he emphasized embedding policies into our practices.
He gave an example of API for Aadhaar data where he did below
a) Ensured that an API call has to be first digitally signed to ensure the authentication (or know the caller).
b) Then, he ensured that API responses were encrypted.
He also mentioned that the DBA of the Aadhaar database cannot see the data. Neither entity can see the data unless anyone goes through an approval process, which will have enough audit trails.
The author of this article is Ranjan Bhasin, one of the ITNext100-2023 winners, he concluded Dr. Pramod Varma's keynote session based on his experience. Ranjan works as Senior Director – Information Technology with Greystar India, where he has set up the IT function and now setting up Automation CoE.
Image Source: Freepik
Add new comment