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.
Balancing Public vs. Private Cloud in the Hybrid Multicloud Era
What are the benefits and differences between private cloud and public clouds? Knowing them can help optimize hybrid and multicloud IT operations.
The classification of cloud computing models into public, private and hybrid and multicloud 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.
Momentum has moved behind hybrid multicloud systems as the ideal model, but when deciding when to use 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.
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 workload or business function. Moving apps, data and workloads between private and public clouds may become less of a burden and more frequent as innovation removes friction between private and public clouds.
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 sign up for and use a web-based interface or Application Programming Interfaces (APIs) to access and use the infrastructure and services.
The defining characteristics of public clouds are their scalability, elasticity and pay-per-use methodology. When it comes to a pay-as-you-go model, companies can set up, scale and tear down servers on a per-use basis, depending on the extent of the compute, network or storage resources they need. There is no obligation to use the hardware, software or ecosystem for a minimum or extended length of time. For predictable workloads, though, customers can also choose to reserve instances for a longer period of time taking advantage of discounts offered by some public cloud companies.
Private Cloud for the Enterprise
A private cloud is typically referred to data centers managed by customers on their own, typically by purchasing hardware and software and making the resources available to their employees and customers within a private network. Some organizations have multiple data centers across different geographic regions and locations that are connected to each other in order to serve constituents in different locations across the world.
The key difference between a private cloud and a public cloud is that in the case of private cloud the resources are only available privately to an organization.
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.
The infrastructure of the private cloud can be set up on-premises in a company’s existing data center 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, with the majority of IT leaders agreeing that hybrid cloud is the ideal model, the choice is not between private or multiple clouds but rather how to make applications and data work across the hybrid ecosystem.
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?
Sometimes a public cloud service may not provide the desired control or performance required by an organization. This is when the private cloud can be used to maintain speed and agility.
Optimum resource utilization: One of the key tenets of public cloud usage is pay-per-use. However, organizations typically end up paying for resources, whether the application is being used or not. Given that public clouds offer standard t-shirt sized compute resources, many organizations often don’t get the exact sizing for resources, resulting in deployments with micro-waste of resources and higher TCO. Therefore, in cases where the workloads are predictable, it is often more cost effective to run them on a private cloud.
Optimize applications: It is often 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 by containerizing and “cloud-enabling,” 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, monitoring and security resources.
In many cases, such as with Nutanix Clusters, the private cloud can be used to develop applications that unify the whole hybrid cloud ecosystem with a ‘single pane of glass’ interface.
Adapt technology according to requirements: The adaptability of private clouds is not just restricted to apps; compute and network. 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, some underlying technologies can be adapted or tweaked 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 a combination of Nutanix Enterprise Cloud, clusters, and other solutions.
“That’s primarily down to the Prism management interface which allows me to 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. 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.
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.
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. It also offers a pay-as-you-go model that enables organizations to pay only for the resources they use. 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 with the private cloud, but the public cloud simply does away with the need for data centers 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.
Skilled staff to manage IT infrastructure: Public clouds attract talented people to work 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 battle-hardened and prepared in the event of a disaster or attempted security hack.
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?
As more IT operations become hybrid with private and multiple public clouds, organizations are eager to write and deploy applications in a portable fashion, according to Madhukar Kumar, vice president of product marketing at Nutanix.
“Write once but run anywhere,” he said.
To help get to that frictionless, mobile state, Kumar said cloud platforms need to be virtualized so that the underlying layers of technologies become invisible. That just might tame the raging debate about when to use private or public clouds now and as needs change over time.
Featured Image by Needpix
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.
© 2021 Nutanix, Inc. All rights reserved. For additional legal information, please go here.
A Bird's Eye View of Today's Cloud Computing Service Delivery Models
The cloud enables on-demand services, platforms and applications that allow businesses to scale up pretty much any function or workload.