CEOs and boards of directors are tasking their CIOs to enable artificial intelligence (AI) within the organization as rapidly as possible. As a result, many IT leaders face a choice: build new infrastructure to create and support AI-powered systems from scratch or find ways to deploy AI while leveraging their current infrastructure investments.
For those enterprises with significant VMware deployments, migrating their virtual workloads to the cloud can provide a nondisruptive path that builds on the IT team’s already-established virtual infrastructure.
“Think about this choice in terms of your own home, imagining your core business applications as the very foundation of your house,” says Ken Bocchino, Group Product Manager at Google Cloud. “Just as you wouldn’t demolish that house to start from scratch to build a new kitchen, you look for ways to expand and modernize the existing VMware infrastructure while preserving the integrity of the original.”
Infrastructure challenges in the AI era
It’s difficult to build the level of infrastructure on-premises that AI requires. The networking, compute, and storage needs — not to mention power and cooling — are significant, and market pressures require the assembly to happen quickly. AI workloads demand flexibility and the ability to scale rapidly. For many organizations, building this capacity on-premises is challenging.
Cloud platforms offer purpose-built infrastructure on demand, but there are IT concerns that this involves expensive refactoring, along with what could be a time-consuming and tricky migration. However, organizations don’t have to build entirely new applications. New functionality, including AI capabilities, can be developed with cloud-native services while remaining interconnected with existing infrastructure elements.
A true hybrid approach
The partnership between Broadcom and Google Cloud provides enterprises with a strategy for maintaining their VMware operational models and integrating cloud-native services. Google Cloud VMware Engine enables enterprise IT to nondisruptively extend their on-prem environments to the cloud and easily run workloads in Google Cloud — without having to make any changes to the architecture. There’s no downtime, and all networking and dependencies are retained — as are other benefits (see this IDC Business Value study). IT teams maintain operational consistency by using their familiar on-premises tools to manage cloud workloads, eliminating retraining needs.
Once migration has occurred, the journey toward AI begins, and it can be broken down into three basic steps — each of which can be a stopping point if the organization’s needs are fully met:
- Front door modernization. Organizations frequently begin by enhancing how users access applications. Users can take advantage of cloud-native load balancing and security capabilities such as Google Cloud Armor, which protects against distributed-denial-of-service (DDoS) attacks and provides a web application firewall (WAF). In this way, IT can employ the cloud’s ubiquitous access and security features without having to refactor and re-network their applications.
- Application layer evolution. The next step often involves rethinking specific elements of the application stack, potentially developing new components that use cloud-native services while maintaining connections to existing systems. But this is not necessary to achieve AI enablement.
- AI and analytics integration. Organizations can enable powerful analytics and AI capabilities by linking VMware-hosted data with services such as BigQuery and Vertex AI. In fact, this can be done without needing to refactor or develop new components, if that’s not something IT wants to undertake.
“With Google Cloud VMware Engine, it’s easy to integrate Google’s Vertex AI directly into the VMware environment,” says Myke Rylance, client solution architect at Broadcom. “It’s like adding high-tech ‘smart rooms’ in your house, where both the old and new structures communicate and work together seamlessly. In Google Cloud, IT has all that it needs to scale up quickly to enable AI with their existing virtual infrastructure.”
Find out how easy it is to incorporate AI into your virtual workloads with Broadcom and Google Cloud. Find more information by clicking here.