Deciding between private and public cloud

cloud computing machine learning autonomousInnovation and technological agility is now at the heart of an organization’s ability to compete.  Companies that rapidly onboard new products and delivery models gain competitive advantage, not by eliminating the risk of business unknowns, but by learning quickly, and fine-tuning based on the experience gathered.

Yet traditional IT infrastructure models hamper an organizations’ ability to deliver the innovation and agility they need to compete. Enter the cloud.

Cloud-based infrastructure is an appealing prospect to address the IT business agility gap, characterized by the following:

  1. Self-service provisioning. Aimed at reducing the time to solution delivery, cloud allows users to choose and deploy resources from a defined menu of options.
  2. Elasticity to match demand.  Pay for what you use, when you use it, and with flexible capacity.
  3. Service-driven business model.  Transparent support, billing, provisioning, etc., allows consumers to focus on the workloads rather than service delivery.

There are many benefits to this approach – often times, cloud or “infrastructure as a service” providers allow users to pay for only what they consume, when they consume it, as well as fast, flexible infrastructure deployment, and low risks related to trial and error for new solutions.

Public cloud or private cloud – which is the right option?

A cloud model can exist either on-premises, as a private cloud, or via public cloud providers.

In fact, the most common model is a mix of private and public clouds.  According to a study published in the RightScale 2015 State of the Cloud Report, enterprises are increasingly adopting a portfolio of clouds, with 82 percent reporting a multi-cloud strategy as compared to 74 percent in 2014.

With that in mind, each workload you deploy (e.g. tier-1 apps, test/dev, etc.) needs to be evaluated to see if it should stay on-premises or be moved offsite.

So what are the tradeoffs to consider when deciding between private and public cloud?  First, let’s take a look at the considerations for keeping data on-premises.

  1. Predictable performance.  When consistent performance is needed to support key business applications, on-premises IT can deliver performance and reliability within tight tolerances.
  2. Data privacy.  It’s certainly possible to lose data from a private environment, but for the most part, on-premises IT is seen as a better choice for controlling highly confidential data.
  3. Governance and control.  The private cloud can be built to guarantee compliance – country restrictions, chain of custody support, or security clearance issues.

Despite these tradeoffs, there are instances in which a public cloud model is ideal, particularly cloud bursting, where an organization experiences temporary demand spikes (seasonal influxes).  The public cloud can also offer an affordable alternative to disaster recovery and backup/archiving.

Is your “private cloud” really a cloud at all?

There are many examples of the same old legacy IT dressed up with a thin veneer of cloud paint.  The fact is, traditional IT’s complexity and inefficiency makes it unsuitable to deliver a true private cloud.

Today, hyperconverged infrastructure is one of the fastest growing segments in the $107B IT infrastructure market, in part because of its ability to enable organizations to deliver a cloud-operating model with on-premises infrastructure.

Hyperconvergence surpasses the traditional IT model by incorporating IT infrastructure and services below the hypervisor onto commodity x86 “building blocks”.  For example, SimpliVity hyperconverged infrastructure is designed to work with any hypervi­sor on any industry-standard x86 server platform. The combined solution provides a single, shared resource pool across the entire IT stack, including built-in data efficiency and data protection, eliminating point products and inefficient siloed IT architectures.

Some of the key characteristics of this approach are:

  • Single vendor for deploying and supporting infrastructure.  Traditional IT requires users to integrate more than a dozen disparate components just to support their virtualized workloads.  This causes slow deployments, finger pointing, performance bottlenecks, and limits how it can be reused for changing workloads. Alternatively, hyperconvergence is architected as a single atomic building block, ready to be deployed when the customer unpacks the solution.
  • The ability to start small and scale out without penalty.  Hyperconvergence eliminates the need for resource allocation guesswork.  Simply start with the resources needed now, then add more, repurpose, or shut down resources with demand—all with minimal effort and cost, and no performance degradation.
  • Designed for self-service provisioning. Hyperconvergence offers the ability to create policies, provision resources, and move workloads, all at the VM-level, without worrying about the underlying physical infrastructure.  Because they are software defined, hyperconverged solutions can also integrate with orchestration and automation tools like VMware vRealize Automation and Cisco UCS Director.
  • Economics of public cloud. By converging all IT infrastructure components below the hypervisor and reducing operating expenses through simplified, VM-centric management, hyperconverged offerings deliver a cost model that closely rivals the public cloud. SimpliVity, for example, is able to deliver a cost-per-VM that is comparable to AWS, including associated operating expenses and labour costs.

It’s clear that the cloud presents a compelling vision of improved IT infrastructure, offering the agility required to support innovation, experimentation and competitive advantage.  For many enterprises, public cloud models are non-starters due to the regulatory, security, performance, and control drawbacks, for others, the public cloud or infrastructure as a service is an ideal way to quickly increase resources.

Hyperconvergence is also helping enterprises increase their business agility by offering all the cloud benefits, without added risks or uncertainty. Today technology underpins competitive advantage and organizations must choose what works best for their business and their applications, making an approach combining public cloud and private cloud built on hyperconverged infrastructure an even more viable solution.

Written by Rich Kucharski, VP Solutions Architecture, SimpliVity.