You’ve heard of the public and private cloud, and probably caught some of the hybrid cloud buzz—but which enterprise cloud solution is best for your business? The answer comes down to your unique needs—from the size of your workload to your specific security & compliance requirements.
Public Cloud: No fuss and hyper-scalable, but weak on customization
The public, or commodity cloud, is where you can go to procure on-demand compute, network and storage resources via a credit card swipe on a self-service portal. You consume as much (or as little) as you want and pay only for what you use. A primary benefit of this out-the-box solution is hyper-scalability. Experiencing a sudden spike in usage? With the public cloud, you can rapidly provision the resources you need and consume them for as long as you need them.
This is an attractive option for businesses looking for a turnkey solution for relatively simple workloads, but it isn’t ideal for organizations with more complex needs (such as strict security and compliance requirements). The public cloud is a multi-tenant environment, meaning you’re sharing hardware with other businesses and consumers.
While hyper-scale is great for the right workloads, the public cloud doesn’t necessarily translate into hyper-performance. Running persistent workloads (e.g. predictable and non-bursty) can also get expensive very quickly in the public cloud. Public cloud also requires serious on-staff dev ops expertise to fine-tune your environments to workloads, as this infrastructure typically doesn’t come with any hands-on guidance or support.
Benefits and features of the public cloud include:
- Utility-like costing – similar to your electric bill, you pay for what you use
- Increased scalability to ensure additional resources are allocated as demand increases
- Greater peace-of-mind since most service providers take extensive redundancy measures due to higher economies of scale
- Highly integrated because many “-as-a-Service” components follow the public cloud model, enabling immediate access from any internet enabled device
Private Cloud: Higher performance, higher level of service
The private cloud is often a better option for larger enterprises with persistent workloads and robust IT infrastructures, as it enables you to fine-tune performance to the needs of the specific workloads you manage. Private cloud is a lot like running your own data centre and managing the hardware in that you can build it to your exact specifications—only without the capital expense or hassle of actually doing any of that.
Each private cloud solution is custom built to your organization’s needs and existing environment. And while you may not be able to dictate the brand of hardware your solution runs on, you can be sure that your solution is deployed on a purpose-built enterprise platform. While public clouds come with commoditized “white box” hardware, your private enterprise cloud solution will typically leverage best-in-class hardware and be available with a variety of performance tiers so you can tailor the performance you want to the investment you make.
Another advantage of the private cloud is the level of support available. Enterprise cloud providers work with your enterprise from soup to nuts, tailoring infrastructure design, implementing to your specs, communicating during implementation and then providing ongoing support during production. While public cloud providers may offer IT support for a price, they’d much prefer to hand you the keys and let you tune the engine (which is great…if you’re a mechanic).
These features and benefits include:
- Higher control of data since the flow of information is controlled entirely by the organization
- Fixed/known infrastructure since it’s smaller economy-of-scale, meaning capacity investments can be directly correlated to departmental needs.
Hybrid Cloud: Customized solution for varied workloads
A hybrid cloud solution consists of a mix of cloud services, managed hosting and colocation, enabling you to leverage the best of all worlds to create a custom solution optimized for each workload. As a federated solution that spans multiple sites, hybrid is also often multi-site and can include redundancy in case of outage or disaster. Bottom line: there’s less risk, better performance, and greater value. Persistent workloads that require high I/O and network throughput or have stringent security & compliance requirements stay on dedicated hardware, while highly elastic or cloud-native apps benefit from the speedy provisioning and hyper scale of capabilities of the public cloud.
A hybrid cloud solution was the answer for Graduate School USA, whose rapid growth and complex IT needs called for an infrastructure refresh. By choosing a hybrid solution, Graduate School gained the flexibility to add and remove components on demand without interruption, meeting all of its scalability and performance requirements.
The bottom line: Hybrid cloud is ideal for organizations looking to migrate to the cloud while utilizing the security of their on premise infrastructure. Below is a list of several enterprise benefits to the Hybrid option unattainable or much more difficult to achieve in Public/Private cloud scenarios:
- Cloud Bursting/Elasticity – Control Your Consumption, Manage Your Costs
Hybrid clouds offer the greatest flexibility while maintaining complete control should an organization need to “burst” IT performance for a designated period of time. Likewise, cloud “elasticity” allows IT administrators to expand and contract cloud consumption at will or automatically to manage costs.
The Hybrid option allows IT personnel to pick and choose which apps can be uploaded to the public cloud while leaving mission critical apps to the secure, private portion. Leveraging the cloud alongside the on-premise datacenter lowers latency requirements, allowing optimal response times between end user input and server output. Some popular examples of when this bursting would need to occur include:
- Retailers needing to burst their e-commerce site during seasonal shopping months
- Increased server capacity during the release of a new game
- Increased demand on government sites for submitting taxes in April
However, cloud bursting comes with its own challenges, especially if done manually where the margin for error is slim.
- Automation – Controlling Your Cloud Consumption
Continuing from the previous point, automating cloud consumption based on demand trends can go a long way in reducing opex since IT administrators will spend less time managing systems. Organizations can decide when they need to boost IT performance by utilizing both the on premise datacentre and the cloud – a process which can be entirely automated based on historical trends.
As demonstrated by the below graphic, customers leveraging the Hybrid scenario can scale out their resources to the service provider’s virtual datacentre to meet the demand needs of the organization.
- Cost-efficient – Consumption-based Model
When cloud bursting is measured and automated, organizations will witness significant reduction in consumed resources from their own datacenter.
Under-utilization of an on-prem server is like excess square footage in the backyard – money and time is being wasted to mow an area of the lawn that never gets used. In the case of an under-utilized server, money is being wasted on hardware that remains idle, yet requires investment to maintain. Pushing the applications residing on that server to the public cloud reduces that expenditure.
Cloud subscriptions are typically a “pay-as-you-go” model, so scaling back cloud resources to on-prem servers during minimal demand windows further reduces the investment.
- Redundancy – Extensive Disaster Recovery Measures
Redundancy is IT’s best shield against any infrastructure failure. Simply put, datacenter redundancy is the duplication of datacenter infrastructure to eliminate single points of failure. This is invaluable to the organization as a healthy datacenter is essential to maintaining workplace continuity – should the datacenter fail, then business comes to a screeching halt.
Redundancy – or disaster recovery (DR) – is commonly undertaken during the development/testing of a new application. If there was no failover for the application, then months of progress could potentially be lost should the server crash. A relatable example would be the best practices of saving Word documents – always save files to multiple hard drives in case one gets corrupted. Likewise, from a productivity standpoint, duplicating that same file to the cloud allows for anywhere access to the file, regardless of which device one uses.
- If you’re considering the cloud, but don’t know which model is right for your organization, then consider the below questions as a starting point for defining your cloud adoption needs:
- What are your plans, if any, for cloud computing?
- What are your business drivers for moving to the cloud?
- Do you use cloud applications like SalesForce.com, Concur, or WorkDay? Are those relationships owned by IT or your business?
- Do you have any of your “.com” sites or home-grown applications hosted in the cloud?