Believe it or not, but sometimes the Cloud is not the answer and actually it might make more sense to do something a bit different! Today we are going to explore some other options for hosting applications and services that might not be very well suited for the Cloud.
Let’s start with understanding what services or workloads might not be suited for Cloud computing. Effectively there are three main workloads that don’t work or don’t scale well in the Cloud. The first is a heavy compute, which requires low-cost processing, a great example of this would be a development environment with an extremely large processor hungry application. The second is appliances which are provided with native hardware. Great examples of these devices are SIP gateways, Firewalls, and Legacy servers. The third and final set of devices are normally bandwidth-intensive or devices that require super-low latency connections.
Now let’s explore if we could pull off the above using Cloud, then compare with alternative solutions. The aim is to provide comparisons to help you make a decision or at least challenge your thinking to create new and exciting ways to solve technical challenges!
Using the Cloud?
The first is a complex problem to solve, we know that data-intensive and processor hungry applications can be super expensive in the Cloud. That being said if the value provided to the business is significant then the costs could make sense. But we stated the environment was for development and thus the value was low. So how could we do this in the Cloud for a more reasonable cost? Well, we have a couple of choices, the first is using spot instances.
A Spot Instance is an unused cloud instance that is available for less than the On-Demand price. Because Spot Instances enable you to request unused cloud instances at steep discounts, you can lower your Cloud costs significantly. The hourly price for a Spot Instance is called a Spot price. The Spot price of each instance type in each Availability Zone is set by the cloud provider and adjusted gradually based on the long-term supply of and demand for Spot Instances. Your Spot Instance runs whenever capacity is available and the maximum price per hour for your request exceeds the Spot price.
Spot Instances are a cost-effective choice if you can be flexible about when your applications run and if your applications can be interrupted. For example, Spot Instances are well-suited for data analysis, batch jobs, background processing, and optional tasks
The second option is to use infrastructure as code and only spin up the environment as required. We discuss this in more detail here. The final option is to use a VPS provider like OVH or Kamatera. Both offer highly cost-effective solutions for development compute.
Alternatives
The question to answer now, what alternatives can we use rather than relying on Cloud computing. The first option and most commonly used is to build your own Cloud! (Remember Clouds are just servers in a data centre!). You rent space in a data centre using a service call co-location. co-location agreements provide power, rack space, and connectivity. As the client, you provide the compute, storage and networking. Most co-location agreements will also include a service called remote hands. This service allows you to get engineers in the data centre to change disks, reboot servers, and pretty much act as your hands remotely! a simple search on google will show you all the providers available.
The benefits of going down the co-location route is that for large compute, low latency connections the cost will likely be lower. However, it is important to note that as a disadvantage you will need to purchase your own hardware, maintain it, and most importantly license it. So be sure to calculate the real cost of the Cloud and co-location to ensure you are implementing the right solution. As discussed above this solution makes perfect sense for hardware appliances that do not have Cloud-enabled replacements.
The second option is to use a service provider like Next2IT whom offer a fully managed data centre service. A service like this allows you to achieve much of the above, but the solution is fully managed reducing the cost to the business and removing the need to purchase your own equipment plus maintaining it.
The third option is to set up something locally. Many offices are designed with The third option is to set up something locally. Many offices are designed with communication/data rooms. It might make sense to purchase the required compute, storage & network, and deploy the platform locally. Things to note on this approach are that power is important. You will want to ensure you have a UPS to keep systems up in the event of power failure.
Conclusion
In conclusion, it’s important to weigh up the choices. It’s highly probable that the correct solution is a hybrid of different backend infrastructure, cloud, on-premises, and co-location. The pros and cons will ultimately allow you to decide which will work best for you.
If you are wondering how best to configure your infrastructure get in touch and our infrastructure experts will be more than happy to provide guidance on the best infrastructure for your business.
