Tech news sites love stories about disruption, and the public cloud has certainly been disruptive. As hyperscale cloud providers grow and develop new technologies, they tend to frame it as a total usurping of older tech, and it makes sense. “Why Private Cloud Will Lose” makes a better headline than “Public Cloud Enjoys Moderate Growth.”
But the fact is, the private cloud is getting stronger — particularly for mission-critical apps — where the private cloud is already the dominant force. With the launch of the next generation converged cloud platform, Symmetry has added automation and orchestration, enhancing the benefits of the private cloud for the full range of enterprise workloads.
History of Cloud Computing
A lot of ink has been spilled debating what the cloud is and when it began. Some will argue that the predecessors of the private cloud go back to the 1950s, when multiple users would share processing time on mainframes. When the IBM Sabre was built in 1960, it showed why private cloud computing would revolutionize business processes. The purpose-built system automated travel reservation booking for American Airlines customers around the world. But although Sabre was more of a pioneering forerunner, it lacked the flexibility of the modern cloud.
The World Wide Web dates to the 1980s, but services were still fairly primitive. Internet bandwidth was low and compute was expensive, meaning that companies generally had to handle heavy computing onsite in a purpose-built network — there wasn’t even a commercial dial-up ISP until 1989. Without the necessary bandwidth, there generally wasn’t a compelling business case for investing in the cloud.
Things started to change in the 1990s as the Internet became a global phenomenon, with increasingly flexible applications for both consumers and businesses. Virtual Private Networks (VPNs) allowed organizations to connect their landscapes together across geographical regions, and computer scientists worked on ways to harness diffuse networks of computers more efficiently.
However, the rapid improvement in computing hardware is the reason why private cloud and public cloud exploded in the 2000s. Compute had become cheap enough, and bandwidth plentiful enough, that providers could develop a range of flexible IT solutions, including SaaS, PaaS and IaaS.
The Hyperscale Cloud
As services like Google, Facebook and Amazon emerged, providers needed infrastructure that could serve a large user base, scattered across the world. This required ever-growing data centers, networked together to function as a whole. With that many users, cloud capacity management became a challenge. Demand could vary dramatically from moment to moment and location to location. The providers needed to address high demand when needed, while controlling costs as much as possible.
Over time, the hyperscale cloud developed to address these needs. In the hyperscale cloud, off-the-shelf white-label hardware is configured in standardized configurations, governed by a powerful virtualization layer. Hyperscale providers can rapidly scale resources and shift demand, in an automated process.
The hyperscale cloud can also be an inexpensive way to run and service. Any single piece of hardware can fail without interrupting the system as a whole, and the inexpensive commercial hardware can be replaced cheaply and easily. And with sophisticated load balancing between data centers, hyperscale cloud providers can control operational costs and increase utilization.
Providers like Amazon used their experience building their own network to outsource cloud services for other companies. They provided flexible environments for tenant testing and development and especially seasonal workloads. Instead of reserving a certain amount of hardware or virtualized compute on an ongoing basis, companies could pay for processing, storage and memory when they need it, then close it down again when they didn’t.
The invention of the hyperscale cloud certainly changed the face of cloud computing, and its effect on the industry has been growing rapidly. Annual hyperscale spending is in the tens of billions, and growing quickly. The number of hyperscale data centers is predicted to grow from 259 in 2015 to 485 in 2020.
Even more strikingly, those 485 data centers are expected to account for 47% of all servers and 53% of traffic in data centers by 2020. Taken out of context, these predictions can contribute to the perception that the private cloud is in decline. However, in reality it’s onsite hosting that’s on its way out, not the private cloud. In many ways, managed private cloud benefits have actually outpaced the public cloud for companies who require guaranteed performance and uptime for mission critical apps that weren’t designed for webscale.
Why Private Cloud: Limits of Hyperscale Providers
Hyperscale cloud computing has been transformational, but its business model has certain drawbacks that are very difficult to overcome. For one thing, public cloud infrastructure emphasizes scalability at the expense of configurability. Demanding enterprise workloads like SAP HANA require extremely high memory and throughput, which make them difficult to correctly provision in environments like AWS or Azure, which allocate compute in large, standardized blocks. This can decrease ROI, and lead to sluggish performance and low productivity.
This issue is exacerbated by the inconsistent performance of the public cloud itself. The public cloud controls costs by cramming as many tenants into the hardware as possible, and hyperscale providers take this model further than anyone. Fluctuations in demand can dramatically decrease performance in public cloud data centers, slowing system response to a crawl.
Across a worldwide hyperscale cloud, where problems in any one part of the network can affect the network as a whole, the problem can become even more severe. That’s why private cloud SLAs guarantee performance (at least for high-end enterprise private cloud services), while public cloud SLAs only guarantee connectivity. It’s simply not possible to provision the way public cloud providers do while keeping performance consistent.
Hyperscale cloud computing also lacks the visibility a private managed cloud can provide. In a hyperscale environment, the provider is aware of the application layer, but the hardware is completely opaque. When your system slows down or a scheduled task fails to complete properly, it’s very difficult to spot where the problem is. The same problem that’s caused by demand surges one day could be a result of a bad node, misconfigured application layer, or networking issue the next.
In a complex production environment where a minor problem in one application can quickly grow to system-wide effects, this is unacceptable. This lack of visibility is also problematic from a security and compliance perspective, as it undermines your ability to adequately audit and supervise sensitive data.
Meanwhile, in the Private Cloud
To understand why the private cloud is crucial for mission-critical use cases, it’s helpful to take a look back into its history. Private cloud hosting is in many ways a more direct descendent of early computing models. Universities, businesses and government organizations would build expensive, purpose-built machines, which would be operated on a time-share system. MSPs would come in weekly to perform routine maintenance, make sure the backups were up to date and resolve any issues.
As the technology developed, businesses started to move to remote, single-tenant environments, operated by 3rd party providers. For complex ERP systems which required carefully configured hardware and specialized knowledge, the benefits of cloud services were substantial. Private cloud providers were more cost-effective than building and running landscapes in-house, and generally possessed expertise that was very hard to find. However, costs were still substantial, and the early private cloud was not easily scalable. Organizations could only use the hardware the MSP had purchased and installed, which means upgrades were expensive and time-consuming.
Over time, however, hardware and operating costs came down. At the same time, enterprise software was undergoing a shift, developing better tools to simplify migration and automate day-to-day administration, security and compliance. Tools like SAP TDMS cut down on time-consuming repetitive tasks in major projects. Hosting companies refined their internal processes as well, further controlling costs and improving ROI on private cloud projects.
Virtual Private Cloud: The Best of Both Worlds
We’ve got to give the hyperscale cloud credit — their hosting model has yielded remarkable technological transformation. Public cloud virtualization and automation were the reason why private cloud was once thought of as a niche solution. But by adopting these technologies, enterprise cloud providers like Symmetry have learned to synthesize the low cost and flexibility of public cloud hosting, with the performance and customer service of the private cloud.
Virtualization came first. Private cloud hosting has historically been tied to discrete units of hardware, with the costs and scalability limits that entails. The development of the virtual private cloud changed that. Private cloud providers could pool resources among multiple tenants like a public cloud, while maintaining a larger reservoir of compute and better control of demand. From the end user’s perspective, the performance is the same, but with dramatically lowered costs and the ability to scale on demand.
Now, private cloud is undergoing a second transformation — automation. Historically, the private cloud has been a mixed blessing for development. Greater security is an obvious benefit, but it’s come at the expense of long wait times at the help desk. This has often led to shadow IT, where developers are going outside authorized channels to fire up their own public cloud compute, which defeats the whole point of keeping development in the managed private cloud.
With automation, private cloud providers are beginning to allow self-service just like the public cloud. Developers can now provision their environments with minimal wait time, while keeping everything secure within the company’s cloud. Symmetry’s next generation cloud is leading the way in this regard. The cloud still enforces internal policies the way the help desk traditionally would, but without the wait.
Private Cloud Hosting is a Complete Solution
All this private cloud innovation is reshaping cloud strategy on a fundamental level. Companies moving to the cloud have tended to use a multi-cloud approach to leverage scalability for testing and development, and stability and performance for production. This has posed some significant challenges. Tenants need to manage multiple vendors, and invest in cloud integration to make all the pieces work together.
Now, many companies are turning to the next generation private cloud as a complete solution. They can find all the cloud benefits they need in one place, while simplifying their vendor portfolio. They never have to deal with multiple vendors pointing fingers at each other. And because one vendor is solely responsible, it’s easier to resolve problems quickly.
A single-cloud environment also eliminates unnecessary complexity. There’s no need for complex cloud integration, and tenants don’t have to worry about how updates in one cloud will affect the rest of the ecosystem.
A single-cloud landscape also greatly aids security and compliance efforts. An enterprise cloud gives you complete visibility at every level of your landscape, and allows you to centralize controls, making it easier to detect and address compliance issues.
A single-cloud environment is also easier to defend against outside threats. Part of this is due to the security of a vendor-operated enterprise data center footprint. Your data is protected by physical security and access control, sophisticated network security, monitoring and other techniques, providing a level of protection public cloud hosting typically can’t match.
Another reason why private cloud hosting is more secure is the reduction in attack surface. With fewer connections between fewer clouds, you have a smaller area to defend against hackers. And with better visibility, the risk of an attacker slipping through undetected is dramatically lower.
Resolving the Private Cloud vs. Public Cloud Debate For Your Company
The cloud has gone through an incredible transformation over the last decades, and cloud transformation will continue to accelerate. As companies continue to increase efficiency, you need a provider who will stay on the cutting edge, while providing traditional dependability and high-touch service. At Symmetry, we pride ourselves on functioning as an extension of your IT team. Whether you’re looking for a partner to design, host and run your SAP HANA cloud, a consultant to help you plan your next upgrade or just someone to fill in temporarily for IT staff, we’re here to help.
Contact us to learn why the private cloud is the right solution for your business.