Multi-Cloud Architecture Steps Into the Spotlight
Organizations seek heterogeneity despite the complexity
Multi-cloud is when an organization uses more than one cloud computing and storage service from providers like AWS, Azure, and/or Google Cloud Platform in a single network architecture. Each of the platforms are unique in their architecture, as well as in the applications they provide and how they support them, resulting in a complex mesh of security, cost, and benefits.
Due to this complexity, many companies find themselves asking if it wouldn’t be simpler to use just one provider? Why would one want to risk security, time, and effort to source and integrate different services from various providers? The answers to these questions lie in the notion that organizations are willing to adopt best-of-breed applications rather than settle for mediocrity from a single provider in the name of uniformity.
Where Have We Seen this Behavior Before?
Multi-compute architecture is nothing new. Major technology vendors, like Microsoft, Oracle, and SAP, have peddled their wares, purporting the benefits of a single integrated architecture to run various applications such as human resources, finance, and procurement. Yet, end-user organizations did not fall for that trap. One such reason is the fit of a certain vendor’s technology for the specific business model of the organization. For example, manufacturing companies adopted SAP’s applications for manufacturing processes while using PeopleSoft for human resources. Similarly, retail companies adopted Oracle’s applications for procurement and supply chain but used Siebel for CRM. At the same time, both of them adopted Microsoft’s technologies because of lower upfront costs for databases and also for business productivity applications.
No doubt, there has always been the benefit of having various applications running on a single integrated technology platform for the purposes of interconnectivity. For example, onboarding customers in a CRM application is beneficial when it is tied to the finance application to ensure the same customers pay for the services in a timely manner without having to separately create the customer records in two different applications. However, many end-user organizations opted to buy applications from a variety of vendors that could best meet their specific needs and chose to manually integrate them using numerous data integration technologies.
The Evolution to Multi-Cloud
Similar to multi-compute architecture adoption, multi-cloud approaches did not happen overnight. Companies began transitioning their systems and applications from their data centers to the cloud a few years ago after buying into the benefit of lower operational costs. The cloud enabled companies across industries to focus on their core businesses rather than deal with IT, which is a necessity but an overhead nevertheless. In the transition process, companies operated in a hybrid cloud mode with part of their applications running on-premises and the remaining in the cloud. Hybrid cloud refers to a cloud computing environment that uses a combination of on-premises computing, private cloud, and public cloud. Multi-cloud is an extension of hybrid cloud, as it merges on-premises operations with services and applications running on multiple cloud environments.
Businesses select certain cloud providers for the applications they already use. For example, companies may choose the Azure cloud because they use Microsoft SQL Server, Dynamics, or Office360 and then the Google Cloud Platform for certain data science applications. And they may also opt to use AWS for S3 storage. All these may be utilized while still running some of their legacy applications in on-premises data centers.
Benefits and Challenges
One of the key reasons multi-cloud platforms have taken off is they allow organizations to customize their technology infrastructure in a way that conforms to the uniqueness of their business processes. The approach enables organizations to capitalize on the benefits of each platform while minimizing their shortcomings. Many multi-cloud environments take it one step further by supporting organizations with governance, risk management, and compliance with regulations.
But there are also challenges with multi-cloud architectures. The main one being that data is siloed across the different service providers, inhibiting organizations from gaining a holistic view of the data needed to run the entire enterprise. Another challenge is security. Each of the cloud service providers provide good security within their own environment, but when the disparate data has to be accessed across the different clouds, it can create security gaps that can expose the data to unintended users.
For the past 20-plus years, data integration has served well in integrating data across multiple systems, like CRM, ERP, and so on, and delivering them to specific destinations like the data warehouse. Given the migration to the cloud, these tools have stepped up to integrate the data not only from on-premises systems but also cloud systems as well, thus enabling hybrid architectures. With the evolution to multi-cloud, these tools now integrate data that is spread across several public and private clouds, and can deliver it to systems that may reside on-premises or in the cloud.
In addition, the tools include strong security capabilities to ensure the data is protected by the policies instituted by the applications and that it can be accessed by the end users only to the level of authentication they have.
With cloud data integration tools in place, organizations can now freely choose where they want to store their data or run their applications — on-premises, in the cloud, a mix of the two in hybrid mode, or across multiple clouds — without worrying about the complex data management that comes with multi-location architecture. This flexibility enables them to choose the best applications that meet their specific business needs without settling for mediocre applications that may come from one service provider.