Instances 'from under the counter', or how to run hard-to-reach AWS EC2 P-type instances?

Amazon Elastic Compute Cloud (EC2) is one of the flagship services of Amazon Web Services, enabling users to create and run virtual machines flexibly and scalable. AWS also provides instances with massive P-type computing power as part of this service, which we will not run in the standard way from the console. So, how do you go about doing this?

In this article, you will learn:

• how to efficiently and quickly purchase AWS EC2 P-type instances,
• how much AWS EC2 P-type instances can cost,
• why their availability is limited.

Why not run an AWS EC2 P-type instance from the console?

Most standard AWS EC2 instances are easily clickable from within the AWS console. However, this is not so easy with the P family, as these are hugely powerful instances using hardware accelerators or coprocessors, which are used, for example, in complex scientific computing, rendering, high-performance databases, or machine learning. The ability to process large amounts of data quickly can be of great importance for some applications. However, before purchasing a P-type EC2 instance, it is worth carefully analyzing your requirements and expectations. AWS rarely allows you to purchase these instances in the standard way for the sake of your bill (P3.16xlarge instances are only available in the console in a few selected regions). While it is quite easy to swallow a bill deviation of a few or a few hundred dollars in case, for example, we forget to turn off some unused service in AWS, inappropriate and unattended use of P instances can be more painful for our wallet. With this family of instances, it is easy to cough up a monthly bill of $60,000. Of course, the total bill is always an individual issue, but the quoted amount and costs of several thousand dollars for a single day's use of an instance is a fairly common occurrence. AWS EC2 P-type instances are available in various configurations, and choosing the right one depends on the specific application or task you plan to run on the instance. In doing so, it is worth noting that higher configurations offer higher performance but are also more expensive.

Why are AWS EC2 P-type instances so expensive?

The prices of EC2 P-type instances can be surprisingly high for some people. However, it is important to understand that these instances offer high computing power and are optimized for processing-intensive applications. The price of an instance depends on the chosen configuration, its duration, and the region in which it is run.

It is worth noting that P-type instances are available at four levels, from P2 to P5, and are differentiated by the graphics cards located "underneath". The top P5 48xlarge instance is supported by an Nvidia h100 graphics card with a market price of around £150,000. Nvidia has produced 200,000 of these cards and 30,000 support instances on AWS. The hourly cost of a P5x 48 large instance is around $100 for each operation hour. When purchasing a P instance, it is,, therefore,, worth analyzing your needs and adjusting your configuration so that you do not overpay for resources that you will not use. As mentioned, these instances use advanced processor and accelerator technologies that provide extremely high computing performance. This, however, comes at a cost to manufacture and maintain these components, which translates into the price of the instance. The high performance of P instances makes them attractive to many companies, research institutions, and cutting-edge projects. This, in turn, generates high demand for these instances, which can lead to reduced availability. Amazon Web Services tries to balance demand and availability, but there is a risk of unavailable instances in a certain region and time.

How do you buy AWS EC2 P instances?

We have already mentioned that the chances of rooting these instances from the console are slim. So, how do you get them up and running? Through AWS support or through a reseller such as Hosters. However, it is worth noting that purchasing EC2 P instances via AWS support can be more complicated, depending on your specific needs and billing history. Sometimes, it may not even be possible to launch this type of instance, as the decision to allocate resources by AWS is preceded by an analysis of the AWS account billing history, the customer's financial capacity, and the AWS sales department's assessment. Purchasing through an AWS reseller, i.e., Hosters, is simpler, as you can bypass the previously mentioned steps and even settle in PLN and receive a Polish invoice. Besides, as an official AWS reseller and partner with AWS Premier Tier Services status, we are very familiar with the procedure for running P instances and have preferential access to them, which increases the chance of acquiring these resources (remember that their availability is limited).

Summary
AWS P-type instances are advanced and powerful tools that can significantly accelerate data processing and computation in compute-intensive projects. However, choosing these instances should be preceded by thoroughly analyzing our needs and financial capabilities. It is also worth paying attention to the optimal configuration to get the best performance-to-price ratio when purchasing them. With the selection of instances and the purchasing process, we will happily assist you as an official AWS reseller and partner with the highest possible status, AWS Premier Tier Services.