How a Multi-cloud Environment Can Ease an Enterprise’s Digital Transformation Journey

There was a time when almost every organisation, whether large or small, had an in-house IT team, primarily tasked with keeping the firm’s IT infrastructure functioning to ensure minimal downtime. While that used to be the target, in practice, most companies had to contend with numerous outages along with deploying skilled manpower in repair and maintenance tasks. Due to the technological limitations of the time, there was barely any in-house innovation of products and services, as skilled manpower was deployed elsewhere.

With the gradual advancement in technologies over the past two decades in areas such as software development, semiconductor design and technology, processor and bandwidth capacity, today, almost every company can claim to be a cloud and AI company where they are finding innovative uses of various cloud-enabled services. It’s not just start-ups and unicorns that are digital natives and leveraging these technologies to grow, but a wide range of large and small organisations in traditional sectors that are feeling an acute need of adopting a cloud-first strategy to innovate and grow.

Cloud adoption has accelerated the pace of digital transformation across organisations like governments, enterprises and start-ups, along with verticals across industry sectors such as BFSI, healthcare, manufacturing and retail, among others. And, this trend will only gather momentum over the next few years as cloud, or more specifically, multi-cloud adoption becomes mainstream with customers. According to market research firm IDC, revenues from India’s public cloud services market—including infrastructure-as-a-service (IaaS), platform-as-a-service (PaaS), and software-as-a-service (SaaS) solutions—totalled $2.2 billion for the first half of 2021 (January-June), with the overall market expected to expand at a CAGR of 24.1 per cent between 2020 and 2025, and reach $10.8 billion.

With the cloud-adoption market expected to increase rapidly over the next few years, an increasing number of organisations are acknowledging multi-cloud to be a foundational pillar necessary for innovation, collaboration, and digital transformation. For instance, three years ago, Divide by Zero—a pioneer in India’s 3D printing space—upgraded to a multi-cloud strategy from the single cloud it had its workload on. It decided to divide its workload between clouds from Amazon Web Services (AWS) and Google Cloud Platform (GCP) when it was implementing an on-demand 3D printing functionality on its web portal, which needed integration between various apps, complex codes, and large-size 3D files. “It started with disaster recovery (backup servers). We had users pushing 3D design files with sizes running into GBs. As a result, we switched to a multi-cloud environment,” says Swapnil Sansare, CEO & Founder of Divide By Zero. GCP handles its large datasets and runs ML and analytics algorithms whereas AWS handles its web services.

“Dependency on a single cloud service poses business continuity risks in case of inadvertent downtime. Multi-cloud offers a more resilient, secure and effective infrastructure,” says Sayed Peerzade, Executive VP & Chief Cloud Officer of Yotta, a managed data centre provider.

Adopting a multi-cloud strategy provides organisations with the ability to better use existing assets and take advantage of newer ways to store, compute, and analyse data. “It eliminates ‘vendor lock-in’-related performance difficulties, restricted alternatives, or expenses arising from employing just one cloud provider. It empowers businesses to modernise existing services, increase agility, and improve business outcomes,” says Bikram Singh Bedi, Managing Director of Google Cloud India Region. Moreover, the adoption of multi-cloud allows organisations to adhere to data regulations without having to build captive data storage units, and still meet its compliance requirements as most cloud platforms ensure regulatory compliances around data storage and protection.

No wonder then that an increasing number of organisations across sectors are investing in a multi-cloud environment. Their cloud plans and adoption, too, have accelerated due to the pandemic, with most considering a multi-cloud strategy to optimise the management of SaaS applications and meet remote work needs. A recent study conducted by hybrid multi-cloud computing service provider Nutanix revealed that 84 per cent of Indian enterprises prefer a hybrid multi-cloud configuration as their ideal operating model, while 58 per cent are expecting to implement such environments within three years.

Challenges Abound

While multi-cloud deployment has become the new norm, a company with little or no know-how of cloud can find itself in a fix when deciding the ideal configuration based on the requirements of workload, data portability, security, and managing different interfaces, let alone its technicalities. Despite these challenges, most enterprises find adopting clouds easy and are already doing it, explains Gaurav Agarwal, Senior Director of Enterprise and Government Sales at VMware India. “What starts with the consumption of one public cloud, results in the adoption of multiple public clouds. That, in combination with the management of an existing private cloud, creates the challenge of managing multiple clouds consistently,” he adds.

When moving to a cloud ecosystem, different companies have varying strategies and motivations. Some move to the public cloud to replace ageing and costly IT infrastructure while others do it to scale for growth and respond agilely to changing market dynamics. This helps improve reliability and scaleability, cuts costs, and brings in new services and products to the market faster. So typically, applications that need to scale on demand make good candidates for cloud integration. “Based on Frost & Sullivan research, companies use an average of 4.5 infrastructure models within their organisation—but they don’t necessarily play well together. Lack of integration (between different clouds) makes it difficult to leverage data and new technology services,” says Ameeta Roy, Senior Director of Solution Architecture at Red Hat, an open source solutions provider. IT and business leaders face the challenge of deploying apps and services in a hybrid environment. Containers, which can mitigate these problems, are a challenge to instal on multi-cloud, she adds.

“The road to the cloud must accommodate a diverse range of workloads across a diverse set of new cloud apps, existing infrastructure, and hybrid models, also known as the multi-cloud strategy,” says Vishal Agrawal, Managing Director of India and SAARC region at Avaya, a business communications software, systems and services provider.

A multi-cloud environment also means multiple cloud providers to deal with. In India, Microsoft Azure, AWS and GCP continue to be the most preferred public clouds, but many companies struggle to choose between the three. However, John Hernandez, Executive VP and GM of Genesys Multicloud Solutions, says that they do not see this as a workload decision, but a function of the company’s overall investment and strategy—though workload decisions could become a consideration in the future. “For example, some companies may have a strategic partnership with Microsoft, which means they will go for Azure, whereas others could be all-in on Amazon (AWS), or Google (GCP). Another consideration would be the availability of cloud services based on region. Companies in China, for instance, may prefer Alibaba Cloud or Baidu.”

A multi-cloud environment also means architectural complexity. Girish Dhanakshirur, IBM Distinguished Engineer & CTO, IBM India Software Labs, simplifies this, “Public cloud catalogues tend to vary in complexity depending upon the choice of as-a-Service used. IaaS offers a lot of flexibility; however, it also increase complexities. PaaS, on the other hand, offers development velocity, but may require a lot of re-write of existing applications. Adopting container-based deployment helps infuse consistency and future-proofs the investment into applications.”

The in-house IT skill requirement poses another critical challenge as current skillsets of employees may not include cloud deployment skills such as DevOps, SRE, Automation, and cloud security. Moreover, chief experience officers (CXOs) who are short on such skills end up facing challenges during implementation. Since there are often different app development teams in an organisation, there exists a wide range of skills across on-premises and cloud. Enterprises look for consistent interoperability between the two so they can unify development practices.

Cloud interoperability has also emerged as an essential element in enterprise cloud strategies as companies seek to mitigate vendor lock-ins, ensure business continuity, and navigate through fluctuating workloads. “At its core, interoperability requires shared processes, APIs, containers and data models across the multi-cloud environment to enable communication between application components. It also requires dynamic discovery of application components, along with real-time data synchronisation,” explains Shivir Chordia, Azure Business Group Lead at Microsoft India.

For instance, Future Generali India Life Insurance (FGILI), had generated a vast amount of data in the past few years, and had adopted a multi-cloud approach. External customer-facing apps that FGILI agents use to onboard customers and issue insurance policies run on Azure, which is scaleable, fast and cost-efficient. Data that is more sensitive such as financial and investment apps, customer records and internal operations is stored on its hyperconverged infrastructure Nutanix Acropolis, which offers more control, security and privacy. Responding quickly to Covid-19, FGILI used US-based IT services firm Veeam’s advanced replication capabilities and direct restore to Azure, which helped its IT team transfer 100 TB of data to Azure in a week. “Going beyond backup, Veeam’s cloud data management solutions have revolutionised our definition of ‘uptime’ ever since the company switched over from legacy solutions. We are no longer constrained by recovery time objectives (RTO), and recovery point objectives (RPO) of hours or days, as we now think in minutes,” says Byju Joseph, CTO of FGILI.

Lastly, another important challenge enterprises encounter is the different interface for every cloud service they opt for. Rakuten’s experience is a prime example. Due to a chip shortage in the industry, Rakuten was not able to meet internal client demand for GPUs, so instead of making clients wait, it moved the GPU-specific workloads to a public cloud. However, it wasn’t easy for them as investing in multi-cloud involves a lot of upfront planning. “Managing a multi-cloud environment from an operations perspective was a challenge. With workloads spread across different cloud vendors, any outage anywhere can impact business. Humanly it is extremely challenging to monitor distributed workloads, monitor uptime, performance, and many things that can go wrong,” says Ashish Nanjiani, Senior Director of Engineering at Rakuten India. The company used Rakuten SixthSense, which combines AI/ML algorithms to monitor machines, and proactively alerts if something is going wrong in the multi-cloud infrastructure. But not all enterprises have access to in-house solutions to manage multiple clouds like Rakuten did.

Multi-Cloud Management

As managing a multi-cloud environment can be a complex task, there are cloud orchestrators which facilitate not just cloud migration and configuration but also take care of optimisation, security, and maintenance. Cloud orchestrators like VMware, Genesys, Veeam, Verint, Yotta, HPE, and even big IT names such as TCS leverage their extensive ecosystem of alliances and partners to support managed multi-cloud services for organisations worldwide, and provide a customised dashboard for easy access.

While opting for a cloud orchestrator involves additional costs, Anil Chawla, Managing Director of Customer Engagement Solutions at Verint India clarifies, “Passing cloud and application ownership to an OEM as managed services will reduce total cost of ownership and help companies manage the application with experts [available at the] right time. The difference in the cost depends on the scale of implementation, and the services they want the OEMs to manage.”

Not surprisingly, the global multi-cloud market is expected to grow at a CAGR of 30.9 per cent from $1.17 billion in 2017 to $4.49 billion in 2022 as per a report by Markets & Markets, a market research firm.

Nidhi Srivastava, VP and Global Head of Google Cloud Business at TCS, says as enterprises tilt towards multi-cloud environments, interfaces, tools and automation too will mature. Meanwhile industry experts advise enterprises embarking on the migration journey to leverage their company resources, skills and IT capabilities, and business needs to overcome the unique challenges they will surely face.

@nidhisingal