29 October 2024

James Moore, VP EMEA Sales, DoIT

To deploy GenAI on a large scale, many organisations are adopting a multi-cloud approach as it allows them to customise their cloud services, access high-quality products, fortify their cloud architecture, and bolster their negotiation position with service providers. With the increasing complexity and resource demands of AI use cases, the costs associated with supporting them — especially the cloud infrastructure for GenAI — are also on the rise. Therefore, for businesses looking to pursue a multi-cloud strategy, it's crucial to consider various factors to ensure the development of a resilient, secure, and scalable infrastructure, whilst seeing a return on investment.

Tailor made cloud

A multi-cloud approach enables organisations to align workloads with the most advanced available technology, such as GenAI, that companies are eager to implement into their operations. However, effectively and securely harnessing Large Language Models (LLMs) within the enterprise requires a holistic reassessment of organisations' IT infrastructure and cloud strategies. With technology continually advancing, a robust multi-cloud strategy will be indispensable for reaping the benefits these new tools offer, enabling customers to evaluate and develop with different providers to understand which best aligns with their needs.

When devising a multi-cloud strategy, assessing a company's culture, DevOps practices, and technology stack is essential. I would advise establishing clear, measurable objectives aligned with broader business goals such as cost optimisation. This evaluation will shape any organisation’s multi-cloud management plan, creating a resilient, long-term blueprint for digital empowerment integrating essential components of people, processes, and technology.

Data security 101

The frequency of cyberattacks has more than doubled since the onset of the pandemic. To ensure continuous system availability, organisations strive to enhance their resilience. One common strategy is to distribute workloads across multiple cloud regions or providers to minimise downtime and ensure business continuity. While a multi-cloud approach can expand the organisation's attack surface, consistent security policies and compliance measures, including encryption, identity and access management (IAM), and security monitoring tools across all cloud environments, can help mitigate the risk of significant damage in the event of a data breach, as data and applications are separated across different services.

Hyperscalers consistently introduce new security solutions to protect data, applications, and workloads. For example, Google Workspace recently announced proactive enhancements to its platform's security, such as mandating two-step verification (2SV) for all Super Admin accounts and incorporating security keys, Google Authenticator prompts, or verification codes through phone calls or text messages. These proactive security measures demonstrate Hyperscalers' commitment to safeguarding sensitive information and ensuring the protection of digital assets across their platforms.

Declouding workload migration

When migrating workloads to the cloud or between cloud platforms, it's important to consider which workloads should be moved. This requires a thorough application assessment to determine which workloads would provide the most value if moved to the cloud or a different cloud environment. For example, public-facing applications with a global reach are well-suited for the public cloud, while others may need an ROI analysis to see if utilising the scalability of the public cloud will be beneficial.

However, during the planning stage, many companies face a skills gap within their teams, which can create challenges during migration. One way to address this is by partnering with companies to provide training and align strategic goals across engineering teams, ensuring that migration goals are successfully achieved.

Lock-in limitations

It's important to remember that regardless of the type of service provider – whether it's a mobile phone company, utility provider, or mortgage firm – being locked into a single cloud provider can pose significant risks. It limits an organisation's flexibility and can lead to increased costs when attempting to switch providers. Therefore, when choosing cloud services, it's essential to carefully assess the potential for vendor lock-in and strive to maintain a level of flexibility that allows for changes in the future, without incurring substantial switching costs.

Adopting a multi-cloud approach can reduce reliance on any single vendor, giving you more control over your operations. Major cloud providers offer pay-as-you-go pricing, allowing you to shut down your environment, export your data and virtual machines, and switch providers at your convenience. Additionally, using open-source cloud computing platforms and tools reduces dependency on proprietary platforms. Open-source technology provides the flexibility to alter source code to meet specific needs, which separates technology decisions from the choice of cloud vendor.

To fully capitalise on the benefits of multi-cloud, it's crucial to prioritise careful planning and optimised architecture. Legacy IT systems and outdated hardware can hamper operations, making it essential to collaborate with a vendor-neutral MSP to identify the best infrastructure mix for complete agility and workload mobility – and ultimately enabling a business to reach its full potential.