The data center industry in India is currently growing at a significant pace. In 2024, the data center market size is estimated to reach $1926M with 2.01 thousand MW and is expected to grow at $4,597.3M with 4.77 thousand MW by 2029, growing at a CAGR of 18.79% during the forecast period 2024-2029.
Yotta Data Services, a leading end-to-end digital transformation provider, launched Shakti Cloud in early 2024 boasting the distinction of being India's first AI-centric cloud. This case study explores the key features of Shakti Cloud and its potential impact on the Indian AI landscape.
About Shakti Cloud
Shakti Cloud, developed by Yotta, is a cloud platform designed to enhance AI capabilities. At its core lies an infrastructure featuring NVIDIA H100 Tensor Core and NVIDIA L40S GPUs, tailored for handling complex AI tasks. Utilising NVIDIA’s NCP super pod architecture, alongside high-speed Infiniband networking and NVMe storage, Shakti Cloud ensures rapid AI performance. Shakti Cloud integrates with the NVIDIA AI Enterprise Solution, ensuring optimal compatibility and heightened efficiency for AI endeavours. It provides a selfservice portal that equips users with end-to-end ML Ops capabilities, simplifying processes from model construction and training to deployment.
Shakti Cloud addresses a wide array of AI challenges encountered by different user groups. For researchers facing high GPU costs, scalability limitations, and software optimisation complexities, Shakti Cloud offers rapid deployment and integrated tools to streamline thread and kernel optimisation. Data scientists and applied researchers often encounter obstacles such as restricted GPU access, debugging complexities, profiling challenges, and a shortage of suitable AI resources. Shakti Cloud accelerates training processes, provides insightful data visualisations, and furnishes user-friendly interfaces for efficient model development.
Shakti Cloud is hosted within Yotta's 700 MW data center campus, offering limitless scalability supported by high-density data center space, robust power infrastructure, efficient cooling systems, and comprehensive connectivity options including public internet and private connectivity. Additionally, it is fortified by layers of cybersecurity protection to ensure the security of data and operations.
Challenges
The emergence of Artificial Intelligence (AI) has brought immense opportunities, but data centers face several challenges in keeping pace with its demands. Despite its potential to revolutionize the Indian AI landscape, Yotta’s Shakti Cloud primary concerns were:
• Performance Focused Design: Optimizing the NVIDIA H100 Tensor Core GPUs for deep learning and complex AI workloads.
• Scalability & Efficiency: Modular designs that allow Yotta to scale their infrastructure efficiently. As their user base grows, they could add more servers to the cloud without needing to completely overhaul the system.
• Reliability & Uptime: Aligning with Yotta focuses on providing a dependable platform for critical AI tasks for which prioritize reliability and uptime.
Solution: An End-to-End AI Supercomputing Platform
ASUS ESC N8-E11 offers a compelling package for building a high-performance and reliable AI cloud platform. Its focus on “Performance Focused Design, Scalability & Efficiency, Reliability & Uptime” making it a strong contender for Yotta’s Shakti Cloud needs.
ESC N8-E11, a 7U dual-socket server is powered by eight NVIDIA H100 Tensor Core GPUs. Is perfectly designed which will help Yotta’s Shakti Cloud to accelerate the development of AI workload. The ASUS ESC N8-E11 is engineered to provide effective cooling and innovative components that deliver thermal efficiency, scalability, and unprecedented performance, resulting in reduced operation cost. Other features are listed below.
• Faster storage, graphics and networking capabilities
• Modular Design with Reduced Cable Usage
• High level of power efficiency
• Accelerated Networking & Security for Advanced Cloud and AI Workloads
• Comprehensive IT infrastructure solutions
Outcome
• The ASUS ESC N8-E11 server emerged as a strong contender for Yotta Shakti Cloud's infrastructure due to its perfect synergy with AI-focused platform. Equipped with the latest NVIDIA HGX H100 GPUs, the ESC N8-E11 delivers exceptional processing power, ideal for the demanding workloads that a user will encounter.
• Faster training times for complex AI models and smoother handling of large datasets. Furthermore, the ESC N8-E11's modular design allows for effortless scaling for user base growth, ensuring the cloud infrastructure can keep pace with increasing demands without needing a complete overhaul.
• This focus on scalability aligns perfectly with Yotta Shakti Cloud's vision of providing a future-proof platform for Indian AI innovation.
