Why the Public vs. Private Cloud Debate Rages On in the Hybrid Multicloud Era

Companies keep moving from private cloud to public to hybrid on account of ever-changing needs. Is there an ideal balance?

By Dipti Parmar

By Dipti Parmar November 18, 2020

The debate among public, private, and hybrid cloud ecosystems is pretty much settled. At least in the enterprise, where it started raging in the first place. Most large organizations – 93% to be precise – now use multiple clouds in their IT infrastructure, with almost all of them having a hybrid cloud deployment.

multicloud

The classification of cloud computing models into public, private and hybrid involves much more than the underlying architectural differences. The infrastructure and services provided – such as Software as a Service (SaaS), Platform as a service (PaaS) and Infrastructure as a Service (IaaS) – can take the form of various virtualized, hyperconverged, software-defined models.

So yes, there is no debate about the use of hybrid multicloud systems, but when it comes down to deciding on a public or private cloud – as part of the hybrid deployment – for a specific operational process or function, there are still many challenges and uncertainties that IT teams struggle to deal with.

It pays to examine the characteristics, benefits, and workings of public and private clouds first, before deciding which of them would be best for a specific business function.

Public Cloud for the Enterprise

A public cloud model delivers IT services – which could be free, subscription-based, or premium – to organizations directly over the internet. These services are maintained and managed by a single vendor or provider, who allocates the central pool of resources to multiple clients on the basis of computing power or access they require.

A single public cloud system consists of a central server that is responsible for controlling multiple other servers as well as the network using a middleware program and various communication protocols. Users in client organizations connect to the internet and then use a web-based interface to access and use the applications or infrastructure they’ve signed up for.

The defining characteristics of public clouds are their scalability and elasticity. Public clouds such as Amazon AWS and Microsoft Azure operate on a utility-based, pay-as-you-go model, where companies can set up, scale and tear down servers on a per-hour basis, depending on the extent of the compute or storage resources they need. There is no contract or obligation to use the hardware, software or ecosystem for a minimum or extended length of time.

Private Cloud for the Enterprise

A private cloud lies at the core of virtualization, separating services and software interfaces from physical devices. All applications are seamlessly available on the cloud, independent of the underlying hardware.

For companies bound by strict data access and storage regulations or those with stringent security requirements, private cloud is the ideal solution. This security is extended to the network – only authorized users can run access-controlled applications through a private VPN or the organization’s internet.

The private cloud constitutes a single-tenant environment, meaning that only one organization (tenant) has access to all the compute and storage resources within that cloud. These resources can be configured and managed in multiple ways.

Finally, the infrastructure of the private cloud can be set up on-premises in a company’s existing datacenter or externally with the cloud service provider, or a combination of both. A private cloud is the ideal environment for virtualization and hyperconvergence, which enable the organization to pool and allocate resources or scale them up and down.

Making the Choice

Either a private or a public cloud (from a specific vendor) could be clearly better for a particular or typical application or workload that the organization operates. There’s no one type of cloud computing that’s right for all companies.

However, for all practical purposes, the choice of public or private cloud is now made in the context of a hybrid environment, as most organizations use multiple cloud systems for their data and application needs.

The deployment and location of the cloud, or the costs and maintainability of on-premises IT infrastructure can also play a significant role in this operational decision.

When is Private Cloud Better?

There are quite a few cases when public cloud just doesn’t offer enough control or performance and the organization has to turn to the private cloud to maintain speed and agility.

Integrate datacenters with cloud services: The immense compute, storage, and processing capacity that on-premises datacenters offer can only be taken advantage of by cloud-native apps and services that are built specifically to integrate with these resources. Public cloud and SaaS apps are more generalized and their APIs and protocols are not built for direct access to the datacenter’s physical hardware.

Optimize applications: It is difficult to extend legacy or existing apps to the cloud. Public clouds follow modern development practices for their environments, which are not necessarily compatible with the organization’s practices. With a private cloud, developers can tweak the organization’s existing apps containerized and cloud-enabled, without the need to start from scratch. When a private cloud is deployed on-premises, workloads requiring low-latency access to enterprise APIs are speeded up.

Further, developers can “refactor” apps and functions that use microservices and communicate with the datacenter using lightweight protocols, as well as isolate and scale resources up or down. Eventually, the number of microservices that constitute in-house apps keeps growing, and management becomes complex, requiring better service discovery, and monitoring and security overheads. This is feasible only in a fully-managed private cloud.

The private cloud also makes it easy to develop applications that unify the whole hybrid cloud ecosystem with a ‘single pane of glass’ interface.

Develop cloud-native applications: Application portability is a central feature of any cloud platform. Cloud-native applications built on the private cloud can be moved across clouds using their tool chains, without compromising on compliance or security compliance.

Adapt technology according to requirements: The adaptability of private clouds is not just restricted to apps; compute, network, and storage systems can also be set up and reconfigured at will according to the changing needs of the organization. Since multiple tenants are not using the same resource pool as in a public cloud, any technology can be adapted or tweaked with a bit of code to conform to specific requirements.

"One of the key benefits from my point of view is a much greater insight into what’s going on inside the box,” said Jeff Longley, Systems Administrator at WD-40, after they virtualized business-critical workloads with the Nutanix Enterprise Cloud. 

“That’s primarily down to the Prism management interface which allows me keep a close eye on everything through a single console. Not that I need to do that very often – it just works.”

Grant granular user access: Public cloud, including IaaS and SaaS platforms, offer “package access” to different features to individual users. Although users log in independently, everyone accesses the same shared resource pool. Private clouds enable separate access to resources on an individual or organization basis.

Meet compliance and security standards: The private cloud offers an isolated network environment that gives full control and visibility into user access, authentication and identity management, as well as data storage, transfer, and management operations with full application-level security.

This is critical for companies in finance, healthcare, and other highly regulated industries – in fact, there are standard data security and privacy practices for Sarbanes Oxley, HIPAA and PCI that can be met via known private cloud implementations. A public cloud with a universal or simplified security model would fall short of meeting compliance requirements or require the use of add-ons that may lead to other complexities.

When is Public Cloud Better?

When the organization can carry out routine operations with a readymade application that can scale, and save costs in the process, they need look no further than the public cloud – it is almost the “default” option in most scenarios.

Save costs: The single biggest advantage of the public cloud is that it offers enormous infrastructure resources without any upfront expenditure for setting it up or deploying; a pay-as-you-go model that enables organizations to pay only for the resources they use. And yes, these resources can be added in miniscule amounts or scaled down when not in use.

Even when the organization is setting up a new project or wants to test out an application, enough resources are immediately available – and can be purchased and separately budgeted for.

Hardware costs are practically zero – the private cloud is a replacement for datacenters, but the public cloud simply does away with the need for datacenters altogether, where applications exist to take care of the workload.

This means that CAPEX (capital expenditure) is fully transferred to real-time OPEX (operational expenditure) – a great way to level the playing field for smaller businesses with less capital at their disposal.

Scale capacity on demand: Handling usage spikes in non-business-critical applications or meeting demand for more resources is a breeze with public cloud systems. Unlike the private cloud, admins don’t need to bother with provisioning or optimal utilization of resources. It’s not a big deal for the providers either, because they have great economies of scale.

Build redundancy and availability within budget: Private clouds are the go-to way to build a highly available, geo-redundant infrastructure. However, true redundancy is beyond the reach of most SMBs as it requires at least two different datacenter locations, which even enterprises are reluctant to budget for.

That said, in today’s hybrid cloud setups, many companies have found a workable alternative in public clouds for disaster recovery (DR). They deploy a “warm DR” infrastructure, which puts the production environment in a private cloud and the recovery environment in the public cloud, which contains a full backup of the production data and applications but remains non-operational until and unless a disaster occurs. This marries the availability of the private cloud with the cost savings of a public cloud.

Have competent staff managing the infrastructure: Public clouds attract the best people working for them – and on their platforms – because standardization of technology provides ample career opportunities for technical people. Staff working for the providers themselves, such as Google and Amazon, are paid top dollar as they are key to the core business.

This means that the admins responsible for maintaining, updating, and managing public clouds are on the whole more knowledgeable – as well as battle-hardened and arguably better prepared in the event of a disaster or attempted security hack – than in-house IT teams employed by an organization to keep the private cloud going, whose competence is at best untested at critical times.

Strategic and Smart Cloud Adoption

Neither public nor private clouds are going away anytime soon. In fact, both are venturing into each other’s territory and there are few effective ways of making apples-to-apples comparisons. As technology evolves and new regulations and security concerns emerge, the battle is intensifying mainly in three spheres – costs, complexity and security.

A smart cloud adoption strategy should be based on outcomes such as better mobility and productivity of staff, business continuity at all times, reduced operational risk and minimum dependence on a particular vendor or provider.

The key to achieving success in all these areas is optimizing and managing costs – data from Flexera shows that organizations waste about 30% of their cloud budget and overshoot even public cloud budgets by 23%. This indicates companies tend to oversimplify their cost considerations without analyzing their needs properly.

The solution is to relentlessly reduce operational complexities and make the infrastructure agile. This can only be done with a defined and documented cloud implementation strategy. For each proposed public or private cloud implementation, the key questions CIOs need to be asking are:

  • How well does the solution address the functional requirement?
  • What are the criteria for provisioning and de-provisioning and what opportunities exist for automation?
  • What is the estimated TCO and how will consumption be monitored?
  • How does the deployment and scale affect organizational and compliance policies?

Dipti Parmar is a marketing consultant and contributing writer to Nutanix. She writes columns on major tech and business publications such as IDG’s CIO.com, CMO.com, Entrepreneur Mag and Inc. Follow her on Twitter @dipTparmar or connect with her on LinkedIn.

© 2020 Nutanix, Inc. All rights reserved. For additional legal information, please go here.