Build software people love at
San Francisco, Dec. 4-7 2017
$87 of free trial credit. No credit card required.
Pay $0.03/GB-hr and let the platform provide everything your app needs to run. Scale out with app instances as your business needs grow. Expand your app with popular 3rd party services in a single click.
Pivotal Cloud Foundry (PCF) customers can run apps on PCF in a datacenter, PCF on a public cloud, or on Pivotal Web Services with enterprise support. Run application instances anywhere with one contract. Learn more about Pivotal Cloud Foundry
Go beyond the basics of what other platforms offer with these free advanced services that automate your operations and keep your app running
HTTP Routing, Traffic and Load Balancing
App Instance Ephemeral Disk 1GB
Free Storage for Your Application Files
Bandwidth In and Out (up to 2TB)
Unified Log Streaming
Active Application Health Management
To make things clearer, let’s go through some examples to understand what we mean when we say “Only pay for what you need.”
While other application hosting providers force you to choose between predetermined sizes of virtual machines or application containers, PWS charges you based on how much application memory the container takes to deploy your app and for how long you use it. PWS provides fine grained control over your app memory to allow you to tune your application’s memory usage to its actual memory needs. Note that this is application compute memory, not disk memory. All users are entitled to up to 2 GB of disk for the application and temporal disk usage. We’ll get into more detail below, but the “TL;DR” is:
1 app instance
1 app instance
1 app instance
How we calculate cost so that you only pay for what you need
If you want more detail, here we go. PWS uses “GB-Hr” as its unit of calculating application costs, and the current rate is $0.03 per GB-Hr. What does this mean exactly? It means that if you run a 1 GB app for 1 hour, the cost will be $0.03. The memory and duration are two parameters you can vary to determine your costs. The more instances you use, the more it costs proportionally. To better understand this, let’s look at some examples:
Java applications are among the largest applications typically running on PWS. The reasons for this are manyfold, but the primary one is that Java applications are deployed with the app server that is bundled during the buildpack deploy process. A fairly typical java app will be 1 GB in size. For the purposes of discussion let’s look at the cost to run this app for 100 hours:
100 hours x 1 GB x $0.03/GB-Hr = $3.00
Now you would not run a production application as a single instance, but you could run it as two instances for redundancy. The cost would be twice the cost or $6.00
100 hours x 1 GB x $0.03/GB-Hr x 2 instances = $6.00
Ruby on Rails (RoR) applications are commonly found and are not as heavyweight as Java applications. A typical RoR app utilizes about 512 MB (or 0.5 GB) to run. The cost to run this RoR application on PWS is:
100 hours x 0.5 GB x $0.03/GB-Hr = $1.50
100 hours x 0.5 GB x $0.03 / 1024 GB-Hr = $1.50
In this case, a lighter runtime also means lighter cost to you. On many other platforms, you would have to pay for the 1 GB container whether you needed it or not.
Node.js applications are prized for their speed and small memory footprint. A typical Node.js app utilizes about 128 MB (or 0.125 GB) to run. So the cost to run this Node.js application on PWS is:
100 hours x 0.125 GB x $0.03/GB-Hr = $0.375
In this case, the smaller runtime benefits are even larger.
From these examples, you might think that PWS might be limited to divisors of 1024 but that’s not the case. For instance, in the RoR app above it’s been determined that due to the size of a computation the applications needs a little more memory to store all the intermediate results and repeated tests reveal that the application works best with 640 MB of application memory. On other platforms, this level of tuning would not be possible, but with PWS it is a CLI command away.