Emil Sayegh

Oct 9, 2015

Dispelling the Myths of Hybrid Hosting

(This article was first published at Network World. Click here to view that post.)

When more than 20 services on the Amazon Web Services platform failed on Sunday, many of the internet’s biggest sites and apps, including Netflix, Tinder, and IMDb, suffered outages over a 6-8 hour period. The culprit behind these outages was AWS’s NoSQL database DynamoDB, which was overwhelmed by metadata queries leading to substantial downtime affecting multiple AWS services (including AWS Elastic Compute Cloud [EC2] and AWS WorkSpaces).

This is not the first time that Amazon’s DynamoDB led to prolonged outages in one of Amazon’s data centers. With problems like this occurring more and more often, why do customers continue to flock to monolithic AWS databases? It’s likely because the level of hype surrounding cloud hosting is astounding, even when there’s a better solution available: hybrid hosting.

Hybrid hosting lets businesses set up their databases on dedicated servers and their front-end web apps on cloud, then tie everything together with a single click. It’s the best way to architect modern apps, which begs the question: why aren’t more businesses running a hybrid hosting infrastructure?

While many in our industry have recognized that hybrid hosting and the hybrid cloud are “the next big things” in hosting, others are intimidated by what they don’t know. Because hybrid cloud adoption is still nascent, there remains a lot of confusion and myths about technology. In fact, there is a lot about the hybrid cloud that even those who have been using it for years wouldn’t know.

Each year, Gartner ranks the most over-hyped technologies on its “hype cycle.” Its most recent list includes hybrid cloud computing on the slope towards disenchantment.

Gartner Hype Cycle

To those familiar with the Gartner Hype Cycle, this is good news indeed. At anticipated trajectory, hybrid cloud computing is headed toward enlightenment, productivity, and commercialization – a journey many other technologies never make. This clearly reinforces that the hybrid cloud is here to stay, and will out-last many of the other technologies on Gartner’s Hype cycle curve.

Overview of the Hybrid Cloud

Hybrid hosting combines dedicated hosting with cloud computing to capture the benefits of both, while spanning both on-premise and off-premise IT infrastructures. This yields a highly-optimized IT infrastructure, which organizations can use to store protected or resource-intensive data, such as data bases, on dedicated servers while simultaneously leveraging the highly scalable computational power of a public cloud for all their front-end applications.

In an optimized hybrid environment, the cloud can grow and expand on demand, while traditional dedicated server resources perform consistent workloads that require higher I/O throughput. A business that can build, deploy, and grow on a hybrid cloud platform on-demand is empowered with the capabilities to plan for expected growth and handle unexpected growth – without sacrificing user experience or security/regulatory compliance.

Hybrid Networking

In spite of its many benefits, hybrid cloud adoption is still nascent; as such, there is a lot of confusion and myths about the technology and its applicability. It’s time to debunk some myths.

Myth: Hybrid cloud is only used for cloud bursting.

When an application running in a private cloud gets a sudden demand for computing capacity, it can “burst” onto a public cloud to handle that spike. This cannot be a reactive measure, though, and it is difficult to run applications on traditional, dedicated servers and then swap that same workload to the cloud at will. For cloud bursting to work properly, applications must be designed from the ground up with that in mind; the vast majority of applications are not built this way. It takes special skill and intent to build applications that know how to “burst” to the cloud.

Hosting on a hybrid infrastructure does not magically make an application cloud burst; the application must be designed for that. Furthermore, the hybrid cloud must allow for the cloud burst at the networking level, which requires integration of hybrid at the networking level. It is unreasonable to expect legacy applications running on traditional dedicated servers, to just swap their workloads to the cloud on demand.

Hybrid cloud cannot, in fact, be used for cloud bursting unless the application was designed for that. Combining an adequately designed application with a hybrid cloud infrastructure, however, would enable an organization to build up an auto-scaling and burst-capable application on hybrid cloud infrastructure.

Easy Drag and DropMyth: Hybrid cloud is complicated to implement.

This is only true if hybrid cloud is done in a non-automated, non-productized manner. If an organization attempts to build its own configuration, things can quickly get complicated and it can take weeks to implement. However, when hybrid cloud technology is implemented through an automated platform, it can be done in less than a few hours, if not minutes.

Ideally, a hosted hybrid cloud solution should be designed with drag-and-drop functionality in mind for every component. This method allows you to configure your infrastructure the way that you want, while keeping the network automation in the background. A drag-and-drop interface makes implementing the hybrid cloud a breeze.

Myth: Hybrid cloud is more expensive.

Hybrid cloud can be much less expensive than a purely dedicated or purely cloud configuration with the proper setup – namely, if cloud servers are leveraged for variable workloads and dedicated servers are leveraged for fixed workloads. There is a possibility for hybrid cloud to run up the costs, but that’s only if bridging devices are used. Done correctly (and without these devices), a business can cut its costs with a hybrid cloud infrastructure.

Take a florist business, for example. Florists are dramatically busier on days like Valentine’s Day than the rest of the year. If a florist pays for dedicated hosting based on those sporadic days, they’re paying way too much. Cloud, too, can be expensive; every other day of the year, the florist sees pretty static load and demand. For the florist to pay for cloud is like staying at a hotel every night instead of buying or renting a home. With hybrid hosting, the florist can run day-to-day business on dedicated servers, add cloud instances for the Valentine’s Day peak, then return to dedicated servers only on February 15. This is the most cost-effective way to go.

Costs can run up when businesses pay premiums for resources that should be fixed commodities like bandwidth or storage. A proper strategy utilizes each element of a hybrid cloud set-up to gain operational and cost advantages.

Debunked!Myth: Hybrid cloud is only for enterprises.

Organizations of all sizes can reap the benefits of hybrid cloud – start-ups and SMBs are even more primed than enterprises to benefit from the hybrid cloud.

Enterprises have the most legacy apps that require a dedicated infrastructure, and can gain a lot from the hybrid cloud by integrating existing environments with new ones. However, changing applications, migrations, and IT approvals can take a long time. Start-ups with fixed workloads (like databases) and variable workloads benefit from the hybrid cloud, too. In fact, we’re seeing more adoption of hybrid amongst startups and SMBs than enterprise due to the agility of decision making in smaller organizations; the hybrid cloud platform is conducive to growth and efficiency, so it is a natural fit for this space.

Because on-demand hybrid hosting is easy to set up and requires minimal configuration, it is ideal for businesses that have small IT teams (especially for IT “teams” that are staffed solo). Hybrid hosting also offers reliability, giving these start-ups and SMBs an edge over their larger, slower competitors.

Myth: Hybrid is good for data redundancy.

“Traditional” hybrid cloud is not good for data redundancy. Consider an organization that stores critical data on its local dedicated server environment. This company runs a redundant system on a public cloud for live failover or immediate data recovery. In the traditional hybrid cloud model, this organization is reliant upon a single physical networking device to bridge the cloud and dedicated infrastructures. The purpose of redundancy is to eliminate points of failure, not add potential network failure scenarios. A traditional hybrid architecture with a “connect” device is a single point of failure.

On-demand hybrid hosting, which is architected at the layer 2 network level to send data through an automated deep layer of networking, is redundant and ensures the FASTEST data transport path from point to point.

In Conclusion

It is clear to all that the hybrid cloud is the future and will dominate IT for the next decade; it is obvious by the repeated AWS failure due to large cloud databases like Dynamo DB. The hybrid cloud is perfectly designed to handle the networking needs of big data as well as the Internet of Things, thanks to its multi-purpose, multi-layered architecture providing superior performance and scalability.

The beauty of enabling the On-Demand Hybrid Cloud on IT infrastructure is that it customizes to the unique needs and usages while optimizing costs. Hybrid is the antithesis to the “one size fits all” approach that companies have had to deal with for years. Figure out what your ideal infrastructural environment is, then make it happen with the hybrid cloud.

Share on FacebookTweet about this on TwitterShare on Google+Share on LinkedIn

Tags: , , , ,