A Vendor-Neutral Approach to Multi-Cloud and Hybrid Cloud Architecture

We recently held our most popular webinar to date on the topic of building a vendor-neutral hybrid and multi-cloud architecture. For those who prefer reading versus watching, I’ll hit some of the major points we covered. Of course, you can just watch the webinar here.

From on-premises to a hybrid cloud strategy

Enterprises are moving the heart of their IT infrastructure away from on-premises environments to empower their digital transformation. Most organizations move to a hybrid cloud infrastructure that is composed of one or more public cloud service providers (AWS, Microsoft Azure, Google Cloud), colocation-based private cloud data centers, and SaaS providers like Salesforce. They may be moving pre-existing VMWare virtualization, or building in a more digital-native fashion. Either way, as IT teams go on this journey, they need to figure out how to choose cloud providers, procure cloud computing and cloud storage, migrate apps and workloads, adapt to cloud automation, and formulate IT resiliency, backup and disaster recovery processes in a cloud environment. Just as in any IT domain, a key concern is how to ensure that cloud vendor selection and rapid cloud deployments don’t lead to lock-in with public cloud providers.

Application context for cloud adoption

It’s always important to understand business context when talking about IT architecture evolution. IT is all about delivering data and applications that power the business, so looking through those two lenses is a helpful place to start. 

Let’s first look at the complexity of the enterprise application landscape and how it impacts the hybrid cloud computing environment. We can get some appreciation for this by examining the average number of custom-built applications in enterprises. According to a 2017 Cloud Security Alliance Report, enterprises average anywhere from 22 to 788 custom applications depending on their organizational size. Those applications are migrating from on-premises data centers or being rebuilt from the ground up as microservices in a IaaS cloud computing environment like amazon web services EC2.

The sheer volume of distributed applications and services moving to the cloud is also impressive. Netskope’s 2019 Cloud Report found that the average enterprise uses 1295 cloud services. Of course, some of these are of limited scope, but some of them are foundational, serve a huge internal and external user base, and integrate with many other apps and services. Think about how Salesforce has become a system of record for so many organizations today. It’s helpful to go beyond public cloud management, and include enterprise SaaS when thinking about hybrid cloud environments, especially where user interfaces and API calls are occurring at scale.

Data management context

Now let’s focus on the data storage side of the story. When enterprises move from on-premises infrastructure to off-premises colocation data centers and hybrid cloud solutions, data decentralizes and becomes inherently distributed. And that occurs while data continues to grow at a high rate. According to the 2020 Seagate Rethink Data Survey conducted by IDC, total enterprise data will grow annually by 42.2% between 2020 and 2022, more than doubling in that time.

Data has gravity, mostly because it’s been so difficult to move. But the problem with the notion of gravity is that data really does need to move to be useful. If you can’t move data effectively to the applications and users that need it, you’ll impede the performance of key use cases.

Adding Dimensionality to “Vendor-Neutral”

Our webinar focused heavily on the connectivity issues related to creating a vendor-neutral cloud platform for IT teams, but clearly there are other dimensions. 

App portability

App portability is a goal that some organizations prioritize.  Some consider Kubernetes as a cloud-native operating system that aids portability because it offers a common control plane for orchestration of containerized apps across different providers and cloud resources. Yet, the real rub for K8s or any other container management or PaaS solution is the degree to which they are integrated with provider-specific computing resources and features.

Multi-cloud

Beyond the matter of app portability lies the choice of multi-cloud as a foundational infrastructure strategy or as a best of breed strategy. The former can be designed to increase provider neutrality but is costly and technically complex to pull off, requiring greater sophistication from devops teams. Best of breed is the more common choice for most enterprises. 

Colocation choices

Another dimension to consider is colocation data center provider choices. It may be tempting to think that just going with a single brand is the way to keep things simple. However, in most cases enterprises find over time that business requirements like geographical placement, service offerings, ability to handle sensitive data or regulatory requirements, and of course cost savings drive towards a multi-colo approach.

The Network is the Computer

Way back in the day, Sun Microsystems had this tagline “The Network is the Computer.” I got teased a bit by my co-presenter about how old I am to recall that vividly (the insolence!), but the fact is that while it’s an old tagline, it’s still absolutely true. In fact, it’s so true that Cloudflare bought the trademark for that tagline and uses it today.

In the webinar, we trace the evolution of the wide area network (WAN), starting from the on-premises reality of internal data centers, HQs and branch offices, tied together via a hub and spoke style architecture using MPLS VPN (virtual private network) services. We looked at how remote access VPNs extended that infrastructure, and then how cloud and SaaS started exerting immense pressures on this legacy architecture by placing their distributed IT resources and network edges within tens of milliseconds of latency of most office locations to optimize user experience over the Internet.

The bending of the corporate WAN to the shape of the cloud created the opportunity for SD-WANs that utilize Internet VPN overlay tunnels to connect branch offices to data centers and SaaS. The success of Internet overlay for branch office connectivity led to the rise of Secure Access Service Edge (SASE) connectivity for remote workers and offices, that folds in authentication, firewall and other security management tools and components into a cloud-based service offering. In the webinar, we unpack the meaning of overlay and underlay, and you can get a deeper treatment on that topic in this blog post.

It’s easy to think that the only enterprise WAN evolution that has occurred is in the last mile, domain of SD-WAN and SASE. Yet a corollary and equally important WAN revolution has rapidly arisen–a new, highly automated middle mile that connects the hybrid cloud core, where the most mission critical workflows and data integrations between cloud, colo, and SaaS serve the heart of the digital business lifecycle. 

The WAN becomes cloud

The middle mile, as called out by telecom research firm Telegeography, is the stretch of the enterprise WAN that occurs between colocation data centers, where the Internet, cloud, SaaS provider networks, and high-speed telco connectivity meet.  What’s notable about the new

middle mile is the degree of automation being applied, making it possible to get high-speed data center interconnection, cloud on-ramps, and multi-cloud routing on demand, with short terms that break the ages old telco model of long provisioning cycles and multi-year, inflexible contracts.

PacketFabric has built the industry’s most scalable, private optical middle mile network offering, with a 50+Tbps backbone, hundreds of colocations data center PoPs and cloud on-ramps, the industry’s only 100Gbps multi-cloud router solution. PacketFabric connectivity can be consumed from 50Mbps up to 100Gbps speeds in minutes via portal and API, with terms from monthly and beyond. Most recently, our backbone VCs offer an even more cloud-like consumption option: hourly baseline and burst (above term) billing, allowing you to support asymmetric data movements for backup and disaster recovery, cloud bursting to meet application and user experience demands, and affordable redundant network connectivity. 

What does all this have to do with creating a vendor-neutral hybrid cloud architecture? It gives enterprise agency, choice, flexibility, agility, and speed to shift and change their cloud architecture on the fly. Teams can optimize IT resources to meet steady state, seasonal, project-based, and unexpected requirements. The WAN has turned into cloud.

Learn more

If you want to understand how the PacketFabric Connectivity Cloud can help you build an agile, vendor-neutral hybrid cloud architecture, you can watch the webinar, explore our services, or simply request a demo and we’ll show you how to build a carrier-class WAN in real-time. If you know you’d like to just get hands on and start building for yourself, register an account and get started now.