Oct 11, 2018

2 Private Cloud Application Models and How to Know Which Is Right for You

Layachi Khodja, Solutions Engineer

When people think about the cloud, what comes to mind is often the public cloud offered by hyperscale providers. This cloud consumption model provides high scalability, flexibility and an on-demand, pay-as-you-go approach. These benefits make the public cloud very attractive.

The private cloud model replicates those benefits in a private or dedicated environment to provide a configuration similar to the public cloud, complete with a notably efficient use of hardware resources. At the same time, private cloud offers SLA-backed service standards, more control over hardware and virtual machines, better control over security and compliance processes, and failover mechanisms.

Not all private clouds are created equal, however, and a plethora of options is available. Let’s start with two major private cloud models.

The Web-Scale Application Model for Private Cloud

This type of application is architected following a web-scale model that does not require a highly available middleware or infrastructure. The application is very scalable and has the built-in intelligence to provide fault-tolerance and mitigate localized disruptions.

For such applications, a simplified scale-out cloud platform is appropriate. Openstack and some similar proprietary cloud offerings are typical examples of this kind of private cloud platform. This type of cloud can be deployed on-premise, in colocation or over a bare metal server offering by an infrastructure provider. It could be self-managed or procured as a service from a cloud management provider.

Openstack does require DevOps and IT expertise to put together the components needed and dispose of otherwise unused parts. As this platform is constantly evolving, some effort is needed to keep the cloud environment up to date and secure. That effort increases dramatically with the complexity of the application deployed and the variety of cloud features used. In case IT staff or expertise is lacking in the organization, such a cloud can be managed by a service provider in-full or in-part, depending on the technical areas that require outside expertise.

Kubernetes is another example. Even though it is not a cloud platform per se, Kubernetes can be used to deploy and manage this type of application with much more ease, albeit with fewer features supported than Openstack.

The Enterprise Application Model for Private Cloud

If the application requires a highly available infrastructure, needs failover support by the underlying platform, makes use of shared SAN storage or requires strict control over processes for security and compliance, then an enterprise-grade cloud platform should be used.

Such applications are typically deployed as managed private cloud environments based on VMware vSphere or Microsoft Hyper-V platforms, with a rich ecosystem of partners and technology providers.

For small and medium-sized companies, this kind of cloud is often procured as a managed service, as its deployment and management call for advanced IT expertise in various technical areas. Additionally, service providers of such clouds are accustomed to IT components and processes needed for the features to support while offering strict control of security, compliance and SLAs. Those providers can be valuable partners and enablers for achieving performance, reliability and compliance.

A Checklist for Selecting the Right Private Cloud Model for Your Business

So how do you know which private cloud is right for you and your applications? Start by looking at your applications’ requirements. Some of the factors to consider are related to the application and workloads themselves, while others are specific to the business structure or IT staff expertise and availability. Here is a list of considerations to help you make your choice:

Scalability

  • Does the application scale out seamlessly with many nodes providing the same service for load balancing?
  • Is data replicated across multiples nodes?

This scaled-out approach is typical of the web-scale applications and cloud model, where nodes are stateless and a failed node can be easily replaced. If this architecture isn’t built in however, then an enterprise application model is more appropriate.

Business Continuity and Disaster Recovery

  • Does your application architecture provide fault-tolerance? In other words, does your application include built-in intelligence to handle failed instances, distribute workloads and move them around seamlessly as needed? If yes, go for a web-scale model; otherwise the cloud enterprise model may provide that layer of reliability.
  • Does the application rely on a highly available middleware that guarantees reliability and fault-tolerance? If yes, the enterprise model may be required.
  • Is quick recovery from a disaster or a service interruption critical? If yes, the enterprise model would be a better fit.
  • Is a very low RPO and RTO required? If yes, an enterprise model would be a strong choice, especially since your managed services provider will likely be able to provide SLA guarantees and offer DRaaS solutions.
  • Is your application capable of tolerating massive disruption in a geographical area with no major impact on service delivery? If yes, a web-scale model would work well, with the application architecture itself being built for it.


Compliance

  • Does the application require compliance with any regulations such as PCI, HIPAA or other standards that call for strict control on security and management processes? If yes, an enterprise model makes sense, especially with the help of an expert managed services provider.

IT Resources

  • Does the enterprise have IT staff with the expertise to manage and maintain the environment and related processes? For a cloud environment, this may include IT, DevOps and even developer teams that might not be considered part of the core business of the organization.
  • Does the organization have an IT security team that guarantees the integrity of the IT environment and protects it against ever-escalating security challenges (e.g., DDoS, intrusion, breaches and data exposure)?

A service provider can be valuable for both models. For the web-scale application, a colo or bare metal provider takes the management and servicing of infrastructure off your plate. The cloud platform itself can be deployed and/or managed by a cloud management provider. For the enterprise model, a service provider offers expertise and extensive resources to manage the platform, handle security and compliance processes, in addition to monitoring resources and engaging other solution partners.

The Value of a Trusted Service Provider

If you are considering an infrastructure refresh to take advantage of the efficiency and scalability that cloud technologies deliver, a trusted service provider like INAP can offer a variety of proven infrastructure solutions to address your technical and organizational requirements. Such offerings range from colocation and bare metal cloud for self-managed private clouds to a fully managed enterprise-grade cloud within a dedicated, private environment. If you’re uncertain, INAP can also help you navigate the complexity of the cloud landscape and guide you toward an optimal solution for your current and evolving infrastructure needs.

Explore HorizonIQ
Bare Metal

LEARN MORE

Stay Connected

About Author

Layachi Khodja

Solutions Engineer

Read More