Ross Brown, Vice President, Cloud Product Marketing at Oracle

Four things customers want from their cloud providers

The cloud computing revolution has left its walled-garden phase in the dust.

Since the early days of the cloud, most customers approached their cloud strategy from the perspective of “which cloud is best for our needs?” But now, as customers are five to eight years into their cloud journey, this question is changing to “Which cloud for which needs?” The demand for interconnected, interoperable cloud services has grown as customers see the limitations of using a single cloud and as public clouds have grown in their differentiation and focus.

Very early on, most companies went with a single cloud – in part because choice was extremely limited. There was also a perception that keeping all workloads in one cloud made for easier management and was less expensive than training staff for multiple clouds.

But today, the fundamentals have drastically changed. There are now several powerful public cloud options, all offering fundamental computing, storage, and networking capabilities combined with a range of differentiated platform services to accelerate innovation. In other words, there is real choice at last. 

This is a great benefit for customers who now realize that different clouds suit different applications based on the design choices of the cloud provider. But there is also a key business rationale for distributing workloads: Very few companies are comfortable trusting 100%, or even 80% of their key applications to a single cloud provider, both because of the inherent reliability/security risks in a single supplier solution, and the growing risk of being trapped on an expensive platform with high costs to integrate. 

The combination of available cloud options and desire to diversify has paved the way for multicloud adoption. And as customers strategize their ongoing cloud commitments, here is a wish list of what they should look for – and expect – from their cloud providers. 

1. Ability to run cloud agnostic applications

The notion of running an application—without code changes—on Oracle Cloud Infrastructure (OCI) or Microsoft Azure, or Google Cloud, or in your own server room is attractive whether it’s for redundancy, failover, backup, or other reasons. As a general rule, the closer you can get to bare metal environments, the more compatibility you have for every application as the constructs of virtual machines, containers and serverless build on that bare metal environment, making the breadth of applications that are portable much broader than just cloud-native apps.

Being able to do so without sacrificing performance or security is a very big and increasingly non-negotiable deal for customers. This is a key characteristic of web-scale companies like Zoom and others that run across multiple cloud providers presently – build for agnostic infrastructure but optimize for each environment.

2. Multiple clouds, multiple strengths

Optimizing for each environment is critical, as the base common denominators of virtual machines, container management and storage are the same across providers. But how each provider designs its environment – shared vs. non-blocking, isolated network models, dedicated bare metal servers vs. shared virtualized servers, flexible vs. static load balancing, etc. – all make a profound difference in performance. These differences especially extend beyond design choices on infrastructure and into platform services: One cloud may be better at powering desktop applications; another could excel at server-side databases, ERP, or CRM software; and another might be best tuned for running webscale workloads. 

Over time, cloud-savvy customers will want to partake of all those services without penalty in terms of security and management capabilities. They will also want to avail themselves of higher-level proprietary software specific to each cloud. These differentiators in both how infrastructure is designed and what advanced platform services each cloud excels at are key drivers beyond the need for multicloud. But to get there, the networks must work gracefully together.

3. Fast, inexpensive data flow between clouds

As companies adopt multiple public clouds for different applications, key data will reside in more than one place. To repurpose that data for tasks running elsewhere, fast, secure physical links between clouds are a must. The Oracle-Microsoft Azure Interconnect is an example of a fast, private, low latency connection between two major public cloud infrastructures that can be a game changer for customers running workloads on OCI and Azure.

Integra LifeSciences, a biomedical company based in Plainsboro, NJ, focuses on making the medical experience seamless and predictable by removing uncertainty for patients. The Oracle-Microsoft Azure Interconnect serves as the backbone of Integra LifeSciences operations, where the company runs Oracle E-Business Suite, Agile PLM, and supply chain analytics and its complete HR systems on OCI, as well Office 365, on Microsoft Azure. Its multicloud strategy has been enhanced via the Oracle-Azure Interconnect in support of moving data and services across OCI and Azure with greater efficiency and reduced latency.

MESTEC, an ISV provider of a manufacturing execution SaaS solution, uses the Oracle-Azure Interconnect to power fast linkage between Oracle Autonomous Database running on OCI as well as Microsoft Power BI and virtual desktops on Azure. In moving its operations from one on-prem data center to deploying parts of workloads in OCI and Azure, there was preliminary concern about latency and performance degradation, according to MESTEC CEO Mark Carleton. But the opposite happened. Instead, “we saw performance improvement across the board in some cases big, in some cases not so big, but no degradation,” Carleton said on Oracle’s Stories from the Cloud podcast.

In addition to requiring fast, secure connections between clouds, many customers need equally fast and secure ways to connect their own data centers with their clouds of choice, which Oracle Fast Connect provides. 

But while nearly all customers acknowledge their desire for extremely fast data transfer, absolutely none of them want to pay egregious fees on data flowing from one cloud to another or out to end users. That’s one reason OCI has long focused on minimizing these so-called data egress charges. Recently Oracle joined the Cloudflare Bandwidth Alliance, a multi-vendor push to minimize those charges.

4. Simple management

As much as businesses seek to maximize the unique strengths of each cloud, they also want to minimize the pain of managing, monitoring, and securing different clouds.

In their ideal world, customers could manage all their IT operations—regardless of where they run—from a single dashboard. Similarly, they want a single place to ride herd on security and identity access controls across venues. OCI’s manageability platform for example, supports the open-source FluentD data collection logs commonly used across on-premises as well as in the cloud. 

The ability to assign access rights to one application running in cloud A and another application running in cloud B based on a user’s verified identity, is a huge advantage in realizing the true benefit of secure multicloud deployment.

That would mean, based on your authorization, you could use one system to securely gain access to Oracle Autonomous Database running in OCI, and your designated applications running in Azure, or elsewhere.

This cloud nirvana is not impossible

Sure, customers are asking for a lot. They want to capitalize on advantages specific to each cloud for key workloads but also run open-source favorites across clouds without fear or favor. They want blazingly fast—yet secure—inter-cloud connectivity without the exorbitant charges that have soured many customers on their cloud providers. And they want to manage all of that multi-site IT from a single dashboard.

It is a tall order, but one that Oracle is pursuing wholeheartedly. Ironically, because OCI is “later” to market than Gen 1 clouds, it is newer and specifically built and designed differently to accommodate the multicloud reality which, frankly, is still seen as anathema to some earlier cloud entrants. 

But the writing is on the wall: Just as consumers abandoned the walled gardens of AOL and Yahoo, business customers will gravitate to providers who treat them like grown-ups and respect and enable their choices.