Private cloud is an essential element of a broader enterprise journey to cloud, including public and hybrid cloud.
Cloud computing has transformed how organizations consume, build, and deliver IT services, enabling the agility and flexibility companies need to adapt quickly to changing business conditions.
Although the term “cloud” is often associated with hosted and public cloud offerings, such as Amazon AWS, Google Cloud, Microsoft Azure, and other smaller service providers, enterprises run more of their applications on private clouds than public due to such issues as cost, compliance, latency, compatibility, and security.
This article provides an introduction to the defining features and benefits of private cloud, while also describing its role as the foundation of a hybrid cloud, which allows companies to capitalize on the advantages of both public and private clouds.
A well-architected private cloud built on hyperconverged infrastructure can provide the agility your business needs, while maintaining the security and control IT requires. It is also the foundation of a holistic cloud strategy, one that incorporates hosted and public clouds into an intelligent and economical hybrid cloud.
First, a quick primer on what “cloud” means. In the broadest sense, cloud computing refers to the delivery of IT services to users on-demand via the internet or a private network. The users consume the services without having to actively manage and maintain the IT resources. This could happen on-premises, in hosted or managed settings, and in other shared, hyperscale public datacenters. The three main categories of cloud computing services are IaaS, PaaS, and SaaS.
IaaS provides the fundamental IT resources that end users require to deploy and run software, such as operating systems and applications. These resources include compute, storage, networks, and APIs, among other things.
The underlying infrastructure, by design, remains largely invisible to the end user. The resources are virtualized and have varying capacities and configurations to meet the needs of the users and workloads.
With PaaS, the execution environment is included on top of the infrastructure elements. PaaS enables users to deploy custom or standardized applications with supported APIs, scripts, programming languages, and tools. End users control their deployed applications and configuration settings, but the providers manage everything below this layer.
SaaS is the next level up from PaaS, where end users consume the application over the network. With SaaS, a provider manages everything up to and including the application itself. Users access the services via an array of devices, such as web browsers and mobile applications. Examples of SaaS include Salesforce and Microsoft Office 365, as well as Desktop as a Service (DaaS). More complex services such as Database as a Service (DBaaS) may be considered either PaaS or SaaS, depending on how they are used.
With public cloud, a third-party service provider delivers services such as infrastructure, platform, and software over the internet for use by the general public. The service providers typically own, operate, and deliver the services using their own data centers.
Customers consume the services on demand, and pay only for the bandwidth, storage, and CPU cycles they consume, along with other services such as monitoring, load balancing, and data transfer. Users typically pay for increments of usage (for example, minutes, hours, or MBs), although some require an up-front commitment, such as minimum contract pricing. Public cloud offers the flexibility to quickly burst, grow, or shrink resource consumption as needed.
With private cloud, an organization provisions services for a single end-user organization to consume. The end-user organization may own and operate the private cloud, or it may pay an external service provider to take on these responsibilities.
Enterprises may also choose to use both approaches at the same time, as you could in an enterprise cloud. The infrastructure may be located on-premises within a business’ datacenter or it may be hosted off-site. Organizations often use private clouds for mission-critical workloads with specific performance, availability, or security requirements. Private clouds offer organizations the versatility of cloud computing while preserving control, efficiency, and security.
Multicloud refers to the increasingly common practice of distributing enterprise workloads across multiple private clouds, public clouds, remote and branch offices, field deployments, and service providers. Organizations distribute workloads across public and private clouds for a variety of reasons, including leveraging the relative advantages of each cloud and reducing the risk of vendor lock-in.
A hybrid cloud combines elements of private and public cloud. Some workloads run on-premises, while others run in a service provider or public cloud (such as GCP, AWS, or Azure) and are portable between them.
Hybrid cloud aspires to seamless interoperability among the clouds, so that the management experience for administrators is essentially the same, whether the workloads are on-prem or in public cloud.
At this point every enterprise has or is developing a cloud strategy. The right strategy for your enterprise depends entirely on the requirements of your business, regulatory laws, and the particularities of your applications, some of which are multitier.
Considerations include a wide array of factors, such as:
These characteristics shape each cloud architecture and drive the decisions about the optimal service for hosting each application tier.
Public cloud serves an array of use cases, which is why companies continue to invest in this approach. However, a recent survey found that many organizations are migrating workloads back from the public cloud to on-premises infrastructure. There are several reasons for this trend.
Organizations can save money by owning rather than renting the resources to support predictable workloads--those applications that organizations use frequently, sometimes around the clock and every day. Examples of predictable workloads include ERP, databases, collaboration apps, VDI, infrastructure services (such as DNS, file, or Active Directory servers), big data, and analytics.
For workloads with elastic usage (highly variable or seasonal), it may make sense to rent the resources in the public cloud rather than purchasing on-premises resources that sit around idle most of the time, or perhaps never get used. Development and testing (known as dev/test) and reporting represent common use cases for this scenario, as they often require spinning up computing and storage resources on demand for specific tasks, and then turning them off.
A key component of any cloud environment - public or private - is being able to monitor resource consumption and allocate it to business initiatives. While public clouds simply send you an overall bill at the end of the month based on your actual consumption, in a private cloud you need to consolidate upfront paid sunk costs when creating spending reports.
It is not enough to account for the hardware and software licensing associated with the underlying infrastructure of a private cloud. You also need to account for the following costs associated with an on-premises datacenter:
To accurately calculate the costs of private cloud services, you need to factor in all of the above to build a true, TCO-based private cloud costing model. Additionally, to drive financial accountability and keep overall cloud costs under control, cloud admins need to create accurate consumption reports for individual users, teams, applications, projects, and more without having to spend a lot of manual effort to do so.
Leveraging an effective tagging strategy and a TCO-based private cloud costing model is a crucial step towards creating accurate and automated chargeback reports that identify spending per cost center in a cloud. With accurate visibility and reporting into spending across public and private clouds, CIOs can better plan their future budgets and drive financial accountability into resource consumption.
Organizations contract with their providers to provide certain baseline levels of service around key metrics, such as performance and availability. Because the public cloud delivers services over the internet, the physical distance and network connectivity in the last mile can cause latency issues that are unacceptable for certain mission-critical workloads.
It's important to consider availability planning for both private and public cloud options, including the potential for downtime or being unable to access resources. With private cloud, you have more control and visibility into the infrastructure, and organizations can tailor their resources to meet their exact requirements, including disaster recovery and business continuity plans.
In a public cloud, security is a shared responsibility. While public clouds take ownership for securing their underlying infrastructure, the end user is responsible for securing their implementation and usage of the provider’s services and resources. A lack of clarity around this arrangement, and a lack of technical domain knowledge, lead to misconfigurations and blind spots that can ultimately leave the end user’s data and applications vulnerable to attack.
This division of responsibility creates additional layers of complexity. Because the end user is free to deploy resources on-demand, it is difficult to establish and enforce a baseline security posture without additional methods for continuous security auditing and compliance enforcement.
It’s often easier to establish a stronger security posture with private clouds. Security administrators have better visibility into resources that users access and deploy as well as greater control over the configurations of the infrastructure delivering the private cloud. Moreover, private cloud enables automated enforcement through built-in network and application security policies.
Compliance has many meanings. For some organizations it's all about government or industry regulations, and for others it’s about ensuring quality, or proof that best practices are being followed. In most cases, the primary goals are around data security. For example, in the U.S., health care providers must abide by HIPAA regulations when managing personal health information, and any company that uses credit card data must follow PCI data security standards.
One issue that can make compliance difficult is the physical location of the data. Regulations often differ among states, countries, and other political boundaries. When using public clouds, customers do not always know where their data resides. This fact can make it hard to determine which regulation applies in a geographic location and whether or not a company is violating specific regulations, such as those against storing citizens’ data beyond their nation’s borders.
Noncompliance with government regulations can result in severe fines and penalties, particularly when the violation causes a breach.
Additionally, though public clouds offer alternatives for many common datacenter services, they all have their own ways of delivering those services, including different configuration options, security policies, and APIs. Organizations lacking expertise in those differences often find it difficult to determine if their cloud services have been deployed and configured in compliance with industry best practices.
It is not uncommon for security breaches to stem from basic configuration errors, such as administrators accidentally leaving their file storage service publicly accessible over the Internet. In all cases, companies must have continuous compliance auditing, as well as swift resolution of audit failures, to mitigate financial and reputational risk.
Private cloud gives organizations better control over security processes as well as issues such as data residency. In general, private cloud administrators are able to address compliance issues more easily due to their familiarity with and knowledge of the entire infrastructure solution.
With the rise of cloud computing, IT is emerging as a strategic partner to the business. Many companies look to private cloud both to increase IT efficiency and to accelerate innovation.
A properly designed private cloud improves IT efficiency by eliminating complexity and leveraging automation to free IT staff from mundane and repeatable tasks. IT staff typically spend the majority of their time on provisioning, updates, data protection, and troubleshooting to address operational requirements and satisfy user requests. Many of these tasks can be automated, including enabling self service for developers and internal users.
Traditional IT environments are hamstrung by silos of dedicated infrastructure around crucial applications such as databases; each silo typically requires dedicated specialists. A private cloud can consolidate all of your mission-critical workloads and data into a single management interface, saving time, effort, and cost, while improving visibility into your overall environment.
Greater efficiency directly supports faster innovation through removing operational friction. A properly designed private cloud helps business innovate by:
Developers and application owners simply can’t do their job when it takes weeks or longer to get the resources they need to develop, run, and change their applications. Developers and testers need fast and easy access to development and test environments, with up-to-date test data, to be productive.
Application owners and DBAs need access to resources to ensure SLAs and end user experience. A private cloud that enables self-service access to these resources can reduce time to market, increasing the cadence at which your company delivers new services and features.
Another bottleneck with traditional IT is deployment. Users can wait weeks or more for new applications to get the IT resources they require to see the light of day. By standardizing infrastructure services and developing with those services in mind, a private cloud can support more rapid deployment and enable CI/CD.
Manual tasks and workflows are the sand in the gears – or worse. It’s nearly impossible to scale application delivery without automation. A private cloud should enable automation so that IT teams can deploy, operate, and scale infrastructure and application stacks with less effort, while providing IT-as-a-Service (ITaaS) to empower development and business teams.
Data is the lifeblood of an organization. The right private cloud should make data more accessible and easier to integrate, accelerating analysis and deepening insight. It also must be able to handle the different types of data and storage that the workloads require and the business demands. Having file, block, and object data under the same roof both increases efficiencies and facilitates the extraction of insights.
Only startups get a clean slate. Established enterprises often have tens or hundreds of traditional business applications that they must continue to support, even as they make new investments in cloud native applications. Most enterprises need a private cloud that can flexibly support both models.
Not all innovation projects are destined to succeed. Failure is a part of innovation. A private cloud makes it easier to resource new development and application projects quickly. Fruitful ideas grow and add resources by scaling out; unsuccessful ideas release their resources and make room for the next effort.
In sum, private cloud increases IT efficiency thanks to simpler operations and greater uptime. It accelerates innovation by providing developers, testers, and app owners with fast access to resources, while clearing the path for new applications to get released. It lowers complexity, which not only reduces IT costs, it also decreases risk of downtime.
A successful private cloud deployment is by no means assured. It must be properly designed. Many existing private clouds suffer from one or more of the following challenges:
Your private cloud may need to adapt to a variety of application needs from traditional enterprise applications to cloud native applications. Inflexible hardware and infrastructure software can make that difficult.
Enterprise environments typically have a need for a private cloud with block and file storage services, in addition to the object storage common in the public cloud. Meeting those needs may require deploying and managing different hardware for each data service, adding cost and complexity.
Separate storage pools decrease overall capacity utilization and limit flexibility. With data playing such a critical role in digital transformation, the importance of addressing this challenge shouldn’t be overlooked.
A private cloud with an inflexible architecture and complex data services inevitably makes automation harder. As a result, it is more time consuming to create automations, automation failures are more likely, and troubleshooting and maintenance is more complex. Moreover, automation tools that are fully integrated into the platform add another layer of complexity.
In a private cloud environment, you may have less visibility of the performance requirements of each workload, and workloads can also change more quickly as users start and stop VMs and applications.
As a result, some workloads can become “noisy neighbors,” taking more than their share of resources and negatively impacting the performance of others. Badly architected private clouds can also introduce single points of failure, bottlenecks, and other flaws that undermine availability.
In a busy and growing private cloud environment, it can be difficult to predict when you’ll run out of performance, especially I/O performance. With multiple types of storage needed, scaling can be disruptive and expensive. The right architecture should simplify scaling; the right tools should simplify monitoring resource consumption while facilitating capacity planning and scenario testing.
The more of these challenges and deficiencies your private cloud is burdened with, the more complex the environment is to manage and the more expensive it becomes in terms of both CapEx and OpEx.
If you’re going to get the return (and the transformation) you are looking for from private cloud, the following features are all required, rather than simply “nice to have”:
A private cloud requires a core infrastructure that enables the advantages of the cloud itself. The complexity of traditional, aka 3-Tier, or server-SAN, IT infrastructure can end up stripping away these advantages. The right hyperconverged infrastructure (HCI) solution provides the essential virtues of cloud computing, including self-healing, simplified capacity planning, easier automation, and reduced management overhead.
There are a number of compelling use cases for choosing public cloud, such as SaaS, development and test, disaster recovery, and data protection. Fortunately the technology has finally arrived that allows you to get the best of both public and private approaches. Hybrid cloud offers the advantages of cloud computing, with the added flexibility of placing your applications and data in the optimal environment.
However, it’s vital that you start with a private cloud that will allow you to build a true hybrid cloud. The ideal private cloud solution includes tools that simplify and unify management of the private and public cloud components of your hybrid cloud environment, including: application and lifecycle management, security governance, and cost governance.
The right private cloud solution should actively facilitate integration with the public cloud. This includes tools for copying, replicating, or migrating VMs and data from one location to another, support for different hypervisors, and the ability to support both VMs and containers.
One of the biggest challenges for any hybrid cloud deployment is that your private cloud and public cloud(s) may have completely different management environments. This makes deployment and ongoing management more difficult and increases the potential for operator errors, security vulnerabilities, and cost overruns. It can be difficult to size equivalent VMs and storage across environments or map functionality in your private cloud to equivalent functionality in a public cloud.
Make sure that the private cloud you build is an on-ramp rather than an obstacle to leveraging the flexibility and agility of a hybrid cloud.
This article has talked about the many advantages that organizations can get from deploying a private cloud powered by a next-generation HCI core. However, even more important than the technology itself is having a deep understanding of your requirements.
Be sure you know exactly what problems you are trying to solve. Technology can’t fix people and process problems. Once you have done a thorough analysis, you will be positioned to build the optimal private cloud for both your immediate business and technical needs, as well as your journey to hybrid cloud.
The ‘cloud’ is one of those fashionable words that has dominated tech discourse as of late,...
Never miss a post.
Subscribe to keep your fingers on the tech pulse.