The multi-cloud conundrum: Strategic choices, challenges, and solutions

According to a Gartner survey dated May 2019, 81% of public cloud users are already working with two or more cloud providers. This is probably far higher now – and the trend is expected to continue – as organizations increasingly seek flexibility, performance, and strategic advantages in their cloud strategies.

There are multiple business reasons for this strategy, the top 5 of which include:

1. Avoiding vendor lock-in: Many organizations adopt a multi-cloud approach to avoid dependency on a single cloud provider. By using multiple providers, they can switch between services and platforms more flexibly, reducing the risk of being locked into a specific vendor.
2. Best-of-breed solutions targeted at specific use cases: Enterprises often choose multi-cloud strategies to leverage specialized services from different providers. Every cloud vendor excels in certain areas, such as AI, analytics, or security. By combining these strengths, organizations can create a more robust and tailored solution.
3. Data sovereignty: With data privacy regulations becoming more stringent, organizations consider data sovereignty as a critical factor. Multi-cloud allows them to store data in specific regions or countries, complying with local laws and regulations.
4. Business agility and innovation: Multi-cloud enables faster innovation by allowing organizations to experiment with different tools and services. It fosters agility, adaptability, and the ability to respond swiftly to changing business needs.
5. Business resilience: Having data and applications distributed across multiple clouds enhances disaster recovery capabilities. If one provider experiences an outage, the organization can switch to another without significant disruption.

This is also a prevailing strategy in the telecommunications sector. A telecom company may implement a generative AI system for optimizing network operations and enhancing customer service. To train the AI models, they would typically use a cloud provider committed to delivering high-performance computing and a suite of AI development tools that process vast amounts of network data efficiently. For the secure storage of customer data and compliance with telecom regulations, they may opt for another cloud provider offering strong capabilities on the security and compliance fronts. For deploying the AI system for real-time network optimization, they may go with a cloud provider who offers an extensive global infrastructure, ensuring seamless operations and low latency across geographies. A multi-cloud ecosystem therefore empowers the company to build competitive applications and services by leveraging the most advanced and cost-effective technologies available.

Challenges and complexities

Despite its many advantages, the multi-cloud journey is often fraught with challenges. Organizations grapple with integration issues, data silos, security concerns, compliance headaches, and escalating costs. The lack of standardization across multi-cloud environments is a major challenge, as is ensuring security and compliance in today’s sophisticated security landscape.

1. Cloud security and risk management: Security concerns are paramount when dealing with multiple cloud providers. Enterprises must ensure consistent security policies across all clouds, manage access controls, and protect sensitive data.
2. Risk assessment: Risks become complex due to varying security postures and threat landscapes across different providers.
3. Complexity and management overhead: Managing multiple clouds introduces complexity. Each provider has its own tools, APIs, and interfaces. Coordinating deployments, monitoring, and troubleshooting across diverse environments can be challenging.
4. Cost optimization and operational overhead: Overheads increase as organizations need skilled personnel to handle multiple platforms effectively. Cost control becomes intricate when dealing with multiple providers. Enterprises must track spending, optimize resource allocation, and avoid unexpected expenses.  Billing transparency across different clouds is essential to prevent cost overruns.
5. Performance and latency: Network latency between different cloud regions or providers can impact application performance. Ensuring consistent user experience across clouds requires careful planning.
6. Integration challenges: Integrating applications and services across multiple clouds can be tricky. Ensuring seamless data flow, API compatibility, and consistent identity management is crucial.
7. Vendor lock-in: Avoidance of lock-ins may lead to custom integrations, which require ongoing maintenance.
8. Compliance mandates: Ensuring compliance with regulations (such as GDPR) across various cloud environments is a challenge.

Throw in a diverse infrastructure mix, multiple cloud vendors, overall governance and observability issues, and deployment and lifecycle management complexities and we have an extremely complicated scenario. For telcos, deploying and management of solutions on remote edge infrastructures is essentially impossible through manual interventions. Without standardization, advanced orchestration, and automation technologies, it can seem formidable to the best of teams working behind the scenes.  

Consider this scenario. For an operator, certain specialized tasks, such as managing GPU-enabled workloads or edge infrastructure, demand specific considerations. For instance, 5G networks rely heavily on gNodeB (gNB) components, which are crucial for handling Radio Access Network (RAN) workloads. These gNBs must support the intensive computational requirements needed for processing and managing data in real time across the network. Content delivery networks (CDNs) must be placed closer to the edge, accelerating the delivery of multimedia content to end-users.

This scenario explains the detailed infrastructure planning required in a multi-cloud environment. In the above example, operators must strategically deploy components across cloud services that specialize in handling high-performance computing tasks and edge computing solutions. This ensures that gNBs efficiently manage RAN workloads, while CDNs deliver content swiftly and reliably to users. The focus is on leveraging the strengths of different cloud providers to optimize performance and user experience.

As-a-service delivery models for a multi-cloud ecosystem

The infrastructure and cloud stack are critical elements of enabling multi-cloud portability and as-a-service delivery models. Cloud-native and containerization of workloads are essential when designing applications with portability in mind. A holistic approach to cloud infrastructure set-up and application design ensures that organizations can leverage the best of what each cloud service has to offer, thereby maximizing their cloud investments while maintaining the flexibility to adapt to future needs.

• Infrastructure as a Service (IaaS): Offers high portability across multi-cloud environments. VM portability, a feature of IaaS, is particularly beneficial for cloudification, allowing businesses to move workloads between clouds.
• Platform as a Service (PaaS): Moderate to high portability. The suitability for multi-cloud depends on how open or proprietary the platform is. CNCF (Cloud Native Computing Foundation) certified platforms, especially those supporting Kubernetes, offer greater portability and are more conducive to multi-cloud strategies.
• Software as a Service (SaaS): Low to moderate portability. The suitability of SaaS for a multi-cloud strategy depends on the specific needs of the business, solution features and the integration capabilities of the SaaS solution with other cloud services. Challenges may also arise in data sovereignty in a dispersed geographical setting.
• Container as a Service (CaaS): High portability. CaaS, particularly services that support Kubernetes, are highly suitable for multi-cloud ecosystems. It provides a consistent environment for deploying, scaling, and managing containers across different cloud providers, enhancing application portability and operational efficiency.
• Function as a Service (FaaS): High portability. FaaS fits well in multi-cloud strategies, especially for event-driven, scalable applications deployed close to end-users or data sources across different clouds.

The role of Kubernetes in this ecosystem cannot be overstated. With the right orchestration tools and automation platforms, Kubernetes can facilitate the management of containerized applications across various cloud environments. The primary challenge remains that not all applications enterprises need are containerized. It is important that the cloud stack can accommodate VM-based applications without adding complexity. Unified orchestration is needed – going beyond simple container management to automated scaling, deployment, backup/DR and lifecycle management of application stacks – making it indispensable for modern cloud-native applications.

Rakuten Cloud: Horizontal cloud designed for portability and multi-cloud support

Rakuten Cloud exemplifies a Kubernetes-based approach that facilitates seamless application portability across various cloud providers, eliminating the need for complex migrations. This approach is especially beneficial for data-intensive or stateful applications, where data mobility and consistent application delivery can be challenging.

Rakuten Cloud's storage and data services ensure uniform application images and workload management across different Kubernetes distributions. It offers capabilities for restoring applications across cloud environments. A unified PaaS interface, operational across all cloud providers, allows applications to move smoothly via APIs on demand, unlocking operational efficiency, agility, and consistent delivery.

Built on the principles of cloud-native approach and hyperautomation, Rakuten Cloud supports cloud-native functions and virtual network functions directly on Kubernetes, bypassing the need for hypervisors. Hyperautomation manages the lifecycle of distributed multi-cloud stacks, optimizing operations for core to edge architectures. Its innovative orchestration platform enables zero-touch provisioning for workloads from bare-metal to VMs across clouds.

By leveraging advanced CI/CD processes and lifecycle management capabilities, Rakuten Cloud simplifies building, provisioning, and managing entire as-a-service application delivery stacks. Challenges related to operations, provisioning, multi-cluster management, and orchestration in a multi-cloud context are effectively addressed, paving the way for streamlined multi-cloud operations.  

Learn more about Rakuten Symphony’s paradigm-changing Telco cloud here.