Hyderabad: The evolving landscape of software architecture is being reshaped by challenges such as integrating siloed data, managing latency and ensuring adaptability in the face of emerging technologies. Prof. Raghu Reddy of IIIT-Hyderabad believes the key to overcoming these challenges lies in developing data-centric and self-adaptive systems. Speaking at the TechForward seminar series, Prof. Reddy highlighted the critical need to move away from traditional application-centric architectures towards enterprise knowledge graphs (EKGs). Unlike conventional systems that rely on isolated data models, EKGs unify fragmented data, facilitating more flexible, dynamic and natural language-driven interactions.This shift is essential to meet the growing demands of internet of things (IoT), augmented reality and virtual reality (AR/VR), and hybrid cloud-edge environments, Prof. Reddy saidAt IIITH’s Software Engineering and Research Centre (SERC), sustainability and adaptability are central to the research, according to Prof. Reddy. He explained, “Collaborating with private organisations, the institute is deploying techniques like federated learning and quantisation-aware training to create energy-efficient systems. These approaches are particularly crucial for addressing the complexities of machine learning (ML) models, such as model drift and large-scale data management.” Prof. Reddy emphasised the importance of hybrid architectures that combine cloud and edge computing. These architectures are necessary for solving challenges such as model compression, security and ensuring real-time responsiveness. He pointed to frameworks like MAPE-K, which allow systems to dynamically adapt to environmental changes, making them more resilient to uncertainties in data and infrastructure.Looking ahead, Prof. Reddy said he envisioned a future of low-code platforms, adaptable designs and sustainable, hybrid systems, where architects play a critical role in navigating rapid technological evolution.
Source link
