How IBM’s India Systems Development Labs enable the AI-First enterprise

Digital transformation is accelerating rapidly, and enterprise players are looking beyond traditional IT applications to intelligent, scalable systems capable of realizing the full value of artificial intelligence. IBM, with its strong heritage of research capacity and enterprise-grade infrastructure, is at the forefront of this shift. Bhavesh Budhabhatti, Director of Enterprise Systems Development of IBM’s India Systems Development Labs (ISDL), is poised to guide that charge—creating next generation computing systems designed for AI workloads in hybrid cloud settings.

In this exclusive conversation, Bhavesh shares insights into how IBM is designing AI-enabled infrastructure, setting new standards in systems performance with new Power11 systems, and helping enterprises address the challenges of deploying AI responsibly and securely, and ready for the future. He discusses IBM’s end-to-end approach to enable enterprises to scale AI safely and productively, in the deep transitions of ever-changing digital transformation. He details aspects of IBM’s strategy, from AI-infused silicon to cloud-enabled end-state visions.

How is IBM innovating to enable businesses to scale AI workloads efficiently in today's fast-changing business landscape?

IBM is helping clients transition to AI-first enterprises by integrating AI into existing infrastructures, workflows, and processes. Our strategy includes AI automation, AI agent orchestration, and trusted AI solutions to enhance productivity at scale. We build AI solutions that integrate the use of AI throughout the entire organization, giving the real technology edge to the businesses. Built on a strong IBM research foundation, we infuse AI solutions across our portfolio offerings right from

(a) Building the IT infrastructure that delivers scalability and optimization,

(b) Portfolio of AI products (watsonx) that accelerate generative AI into core workflows

(c) AI consulting to redesign the end-to-end workflows to scale the AI faster.

For example, we design the AI capabilities into our microprocessors (MMA-Matrix Math Assist unit in Power 10/11, AI accelerator on Telum) that we use to build our IBM Systems, provide a distinct advantage in delivering hardware solutions that accelerates AI workloads, improves energy efficiency, and enables the deployment of AI models across cloud, on-premises, and edge environments. IBM’s Granite AI models are designed to deliver enterprise-grade AI performance efficiently, offering results comparable to larger models while optimizing for business use cases.

How does Power11 respond to the growing needs of data-intensive workloads, including real-time analytics and machine learning?

Power systems are known for delivering high-performance, scalable, and reliable computing solutions for enterprise workloads. While Power 11 extends that lineage forward, with the inherent architectural advantages, it leaps forward in handling the data-intensive workloads. Power11’s AI acceleration and high-bandwidth architecture, combined with IBM’s AI governance tools like watsonx.governance, provide enterprises with the ability to scale AI workloads while ensuring responsible AI deployment.

Some of the key capabilities include  

- Built-in AI accelerators (MMA) to speed up AI workloads.

- High Bandwidth and Low Latency, which is essential for handling the large datasets typically used in AI and ML. e.g.

· 2 Terabyte / Sec bandwidth (Across all interfaces) to the Power 11 chip.

· Significantly higher memory bandwidth over the Power 10 systems, which was already a strong industry-level benchmark.

- Scalable Architecture to handle increased demands as the AI/ML models.

- Extensive parallel processing to process large datasets in near real-time.

- Tight integration of AI-centric software stack, including popular deep learning frameworks like TensorFlow, PyTorch, and Caffe.

Also, more importantly, Power systems offer supreme reliability with 99.9999% or greater availability rating. (As per the ITIC survey of 1,900 C-level executives across 37 vertical markets)

Can you provide more information on the security features integrated into it and how they safeguard sensitive information in a hybrid cloud setup?

IBM Power systems follow a zero-trust security model, ensuring end-to-end protection across hybrid cloud environments. With watsonx.governance, businesses can monitor AI decision-making, track data provenance, and meet compliance requirements globally. This multilayered approach includes Hardware, Operating system, Firmware, IBM® PowerSC 2.0 technology & Hypervisor.

Hybrid cloud environments bring some unique considerations to protect sensitive information and the need for consistent security policies across both on-premises and cloud environments. Power systems safeguard sensitive data in a hybrid cloud by leveraging several key security features

- Real-Time Memory Encryption (hardware-based): Data is encrypted in memory, securing the sensitive information during processing.

- Secure Boot and Trusted Execution: Verify the integrity of the system during boot-up and isolate sensitive workloads in trusted execution environments.

- Confidential Computing: Enable secure enclaves to process sensitive data while ensuring it's encrypted and protected even when in use.

- Role-Based Access Control (RBAC): Fine-grained access control to ensure that only authorized users and applications can access sensitive data and resources.

In general, IBM's vertically integrated design approach makes Power systems highly secure, where both the hardware and software are tightly designed, developed, and optimized together.

What is IBM's approach to maintaining backward compatibility with current Power systems and applications?

IBM Power systems are designed with keeping the backward compatibility in mind with both hardware and software, allowing current Power systems to support applications from previous generations of hardware.

Let’s look at the compatibility from different dimensions.

- Processor Support (ISA- Instruction Set Architecture): Power 11 Systems hardware can still run workloads designed for older processors like Power 7- Power P10, ensuring smooth transitions for hardware upgrades.

- Operating Systems: AIX, IBM i, and Linux (on Power) running on newer Power hardware are backward-compatible. Applications and workloads designed for previous versions typically run without modification.

- Application Support: Applications built on older Power systems can run on newer hardware with minimal changes. IBM ensures that APIs and system libraries used by older applications are supported in the newer environments.

- IBM also offers tools to assist with migrating legacy applications to current Power systems. These tools help identify potential compatibility issues and offer solutions to ensure smooth migration without the need for major rewrites.

In summary, we ensure that businesses continue to run smoothly as they transition to the latest hardware while keeping their existing applications and systems running efficiently. Whether through hardware, software, or migration tools, IBM's approach helps mitigate risks and disruptions associated with upgrading infrastructure while maintaining business continuity.

What is IBM's vision for the future of hybrid cloud, and how will Power systems fit into it?

IBM’s approach to hybrid cloud focusses on providing a computing environment that seamlessly integrates and manages workloads across public, private, and on-premises infrastructure. Our vision emphasizes security, data privacy, and compliance while incorporating AI, edge computing, and quantum computing to drive innovation.

We place a greater priority on advancing hybrid cloud strategy while designing our systems. Power systems enable seamless integration between on-premises data centers and the public cloud. Here are some of the essential functions that make them most suitable for IBM's hybrid cloud strategy

- Optimized performance for enterprise workloads: Specialized processors and capabilities that excel in handling high-performance computing tasks.

- Migration: Move apps and data between private and public clouds with ease.

- Cloud native applications: Build and scale cloud-native apps with containers, Kubernetes, and Red Hat OpenShift.

- Modernization: Leverage existing investment in infrastructure & data and fuel innovations with newer technologies such as containers, microservices, and AI.

With tight integration of Power systems into IBM & other cloud offerings, businesses can leverage best-suited infrastructure for running mission-critical workloads seamlessly across hybrid cloud environments.  For example, RISE with SAP on IBM Power Virtual Server accelerates transformation with SAP S/4HANA Cloud. IBM realized a 30% reduction in infrastructure costs in RISE with SAP modernization project supporting 150,000 ERP users.

What are the greatest challenges that companies face when deploying AI solutions in a hybrid cloud platform, and how does IBM resolve these challenges?

Deploying AI solutions in a hybrid cloud platform can be highly challenging due to the data integration across different environments (on-premises, private/public cloud), handling multi-cloud environments with different interfaces/tools/capabilities, Data movement Latency for certain AI applications, ensuring data security and regulatory compliance etc.

IBM's business strategy is centered around hybrid cloud computing and Artificial Intelligence (AI). IBM helps businesses overcome AI deployment challenges by offering AI-powered automation solutions, including watsonx Orchestrate, to simplify AI adoption and optimize hybrid cloud workflows.

We strongly believe in implementing the right hybrid cloud architecture with intention is one of the most critical business decisions in any organization’s AI journey.  We have built technologies and solutions that can protect the investment in IT infrastructure and maximize ROI with IBM’s cost-effective hybrid platform that’s open and secure.  

Some of the key solutions that we offer are:

- AI-ready infrastructure: Our Servers, storage and software designed to run mission-critical apps with resiliency, security and performance.

- Modernize & migrate applications with AI: Containerize legacy applications and accelerate the time-to-value of hybrid cloud environments.

- Enable an open hybrid cloud to run AI anywhere: A platform-centric approach with Redhat solutions for AI investments to scale with security and consistency.

- Streamline modernization with GenAI: Accelerate application modernization.

We continue to innovate our hybrid cloud offering to enable seamless, scalable AI model development, training, and deployment across on-premises and cloud environments.

What would you tell companies interested in starting their AI and hybrid cloud journey?

While Hybrid cloud and AI technologies are at different stages of evolution, the landscape continues to evolve very fast, and I believe flexibility and scalability of hybrid cloud environments make it the ideal infrastructure for AI applications.

Someone starting a hybrid cloud journey now may have some advantages as cloud computing technologies have matured to a greater extent, there are some proven best practices, and we have better integration capabilities. However, going on this path demands strategic planning, starting from building the right team to making specific choices on Infrastructure, platform, and data management aspects.

While AI technologies have seen massive growth in recent years, they remain in a stage of rapid evolution. At the same time, regulations and compliance around AI covering the aspects of legal, ethical, and responsible deployment are becoming increasingly important. This requires an AI solution to cover and provide agility to all layers, right from data management to model deployment and ongoing monitoring.

So, the most important decision for companies is going to be the selection of Technology providers [for hybrid cloud and AI] who can offer end-to-end solutions, ensure seamless integration, scalability, and security, so they can accelerate AI adoption in their core business without worrying about the complexities of IT implementation, maintenance and regulatory compliances.

What gets you most excited about the future of AI and hybrid cloud, and the role that IBM is playing to drive that?

I am truly enthusiastic about the potential that AI technology holds to shape our industry and impact the world on a larger scale. In today's interconnected world, data is being generated at an overwhelming speed & volume, and AI deployed in hybrid cloud environment, seems to be the right answer to transform these raw data into actionable knowledge, improving decision-making and unlocking new possibilities.

IBM’s leadership in AI is driven by our commitment to open, trusted, and scalable AI solutions. I deeply value the opportunities to contribute to IBM’s AI and hybrid cloud strategy that empowers enterprise businesses, which is deep rooted in industry expertise, powered by home grown AI ready infrastructures & AI platforms (like watsonx) and driven by cutting-edge research.

I am particularly excited about how we at ISDL (India System Development Lab) build and innovate the technologies that are at the heart of IBM’s AI and hybrid cloud solutions. As an example, we are developing IBM Power processor technology, which is the foundation for IBM Power system servers, and these servers are at the forefront of AI and hybrid cloud technologies.