If companies know the differences between the standard hyperscale and the customizable solutions from local cloud providers, they can place workloads where they are best-taken care of. This gives users complete flexibility.
AWS, Google, Azure, and, more recently, Alibaba’s hyperscale’ offerings are all very similar. They offer scalable computing power in a multi-tenant environment that is paid for according to consumption. The offer of the Hyperscale is mainly based on three pillars: flexibility, additional services, and scalability. However, customers pay a surcharge for a high level of flexibility. This is mainly due to the size necessary to cover peaks to provide almost any computing power.
This flexible basis of storage and computing power is used to offer additional PaaS (Platform as a Service) services. These services can be booked quickly without having to develop them yourself, such as scalable databases or payment services. Again, the costs are usually higher than if you were to develop these services yourself.
Hyperscale: Impact Of Data Protection Regulations
Hyperscale are globally present and have data centers in many countries under the umbrella of a global parent company. In the case of US hyperscale, the parent company’s business location is sometimes an issue. The possibility of access by US authorities is possible in the case of US hyperscale. Even a subsidiary owned by a US company is required to allow US authorities access following the parent company’s legislation.
High Entry Price Without An Exit Strategy
Due to the standardization based on the size of the hyperscale, special requests or adjustments are usually not possible. Customers have to adhere to the standards, the way they work, and the solutions that the hyperscale provide. Using these proprietary standards, the use of the offers of hyperscale often leads to unwanted manufacturer loyalty. These define how to do business with them and how to use their infrastructure. For example, AWS S3 has almost become the industry standard, and it isn’t easy to find a way to work outside of AWS when that standard needs to be used.
Smaller Cloud Providers Have A Different Business Model
Compared to the hyperscale, the cooperation with smaller providers, who can undoubtedly operate globally, offers many advantages. These providers do not have to provide a massively scaled overhead to cover short-term high-demand peaks. They are not interested in use cases that would increase their highest peak demand but would result in idle hardware in the remaining time.
In fact, in practice, only a tiny percentage of workloads have high peaks. Most workloads are reasonably stable and predictable. This is a central factor in the business model of smaller cloud providers: this overhead does not exist and therefore does not have to be factored into their pricing. Smaller providers have a completely different business model compared to hyperscale! Their offerings for predictable workloads are therefore much cheaper – in reality, 30 to 50 percent savings than hyperscale – and in some cases even more.
Hyperscale: Paid Support In The Public Cloud
In addition to the higher costs, the offers of hyperscale have other disadvantages. For example, support is chargeable and not included. Customers must purchase a support package that specifies what type of support they will receive. For smaller providers, essential support is usually included in the price. The hurdle for support is much lower, and customers can speak personally to the people who operate their platform. You can get advice, review business models together, or decide what would work best. These things are impossible with the anonymous hyper-scale approach.
Hyperscale: Consider Differences In Cloud Portfolios
Compared to the hyperscale, smaller providers lack the amount of PaaS services. For example, AWS currently offers more than 150 different PaaS services on its platform. If you look at the portfolios, the smaller providers offer everything below the PaaS line. That is, bare metal and dedicated compute, which hyper scalers only offer to a minimal extent. For example, AWS only has a handful of different server models in its portfolio that can be used as bare metal.
In comparison, smaller vendors potentially have hundreds of different models and specifications tailored to the optimal performance. Small cloud providers are also able to add a customizable cloud layer to this infrastructure layer. This enables customers to use a generic service and build their services in a better way.
Smaller providers will most likely not offer a highly scalable database. But they will offer a platform that is cost-effective, stable, and customizable. This allows a company to run its database much better, and customers are not tied to a particular provider and their specifications for how a database should work. Of course, there is a threshold to the use of any service, after which it is much easier and better to build your service than to give a hyperscale a ton of money for a generic solution. But how do you correctly decide where a workload should ideally run?
Hybrid Cloud: Complete Flexibility At Manageable Costs
Hybrid cloud computing has been around for more than 20 years and has become the standard for most cloud infrastructures. Hyperscale and smaller vendors alike have noticed that customers need both pieces and have established connections for the two clouds. The ideal is to combine everything with everything that should be the basis for every portfolio of smaller providers. This is vastly different from hyperscale, who are not interested in letting their customers choose what any competitor has to offer.
Put Workloads Where They’re Best-Taken Care Of
The decision-making process of choosing and setting up the right provider starts with the required workload. Workloads that are volatile and require a lot of flexibility will likely require the performance of a hyperscale. Anything predictable can be run much more efficiently in a smaller vendor data center, tailored precisely to needs. In this way, companies get the best of both worlds and generally do not choose one of the two alternatives.
Public Cloud Of Hyperscale – A One-Way Street
The “cloud-first” approach is generally not a good idea. It is more efficient to look closely at workloads and then find out what the best solution is. There are hardly any customization options with a hyperscale and rarely make economic sense. There’s no going back with cloud-first, while it’s easy to go the opposite path from bare metal to a different cloud environment that can be customized for any specific purpose. Companies should keep their options open by choosing a hybrid cloud model. This gives you the same flexibility at manageable costs and a good price/performance ratio.