Compute Engine

Live Vertical Scaling

Resource management in real time

Scaling up without rebooting — flexible and cost-efficient

Compute Engine / IaaS

Live Vertical Scaling to absorb traffic peaks

You need to be able to alter the resource configuration of your virtual data center on the fly in order to respond quickly to changes such as traffic peaks. Our Live Vertical Scaling helps you do just that. Live Vertical Scaling uses hot add or hot plug procedures to let you boost resources like RAM and the number of cores without rebooting the system. This way you can utilize the full flexibility of the cloud while ensuring optimized performance at an unbeatable price.

Finger pointing to a illuminated display of graphs and data

Scenario in a physical data center

What happens when there is an unexpected surge in traffic on your website? Your servers grind to a halt, your customers are unable to connect, and your revenues take a tumble. This is a real predicament if you are running a conventional data center: you will be forced to shut down the server, at least briefly, to increase the resources at your disposal. In the meantime, a backlog piles up that could easily overload the upgraded server when it is powered back up. Compute Engine

with Live Vertical Scaling puts an end to this problem in your virtual data center.

Icon: Dedicated resources: Server rack with checkmarks

Vertical scaling of RAM and CPUs at the push of a button

All it takes to add resources to your server in the Compute Engine

is operating a slider. The Data Center Designers (DCD) graphic interface lets you increase cores and RAM in very small increments.

This is possible because you are working inside a Software Defined Data Center (SDDC). The capacities for your virtual data center are sourced from within the physical data center’s hardware pool. This means that we can deploy new capacities in real time (software-controlled), whenever you need them. This procedure allows vertical scaling on the fly, i.e. adding new CPU cores and more RAM. You can add up to 62 cores and 240 GB of RAM (AMD CPUs) for each VM, or 27 cores and 120 GB RAM for Intel CPUs.

Scale up — vertical scaling for quick availability

Scaling up, i.e. the vertical addition of server resources, works smoothly irrespective of which operating system you happen to be using. MySQL registers the new capacities without requiring modification. In other words, you can immediately benefit from the boost in performance. This is an advantage over scale out, i.e. horizontal scaling by adding more server instances: horizontally scaled infrastructures are more complex, sometimes require extensive and therefore costly code modifications, and can’t increase performance in a fully linear fashion.

We provide you with a free Ubuntu-based operating system image to make sure that the convenient server management also works without a hitch on your virtual servers in the Compute Engine

. On Linux servers, it is possible to use ACPI (Advanced Configuration and Power Interface) events to signal to the system during live operation that new processor cores were added. This means that the system does not have to be rebooted — physically or virtually.

Cutting costs with Live Vertical Scaling: pay-per-use

Although high performance even during load peaks is indispensable, business owners still need to keep an eye on the costs. Our pricing system is just as flexible as our server resources: with Compute Engine, you only pay for the services that you actually use. That’s why there are no long-term contracts or prefab packages with capacities that exceed your actual requirements. Our IT resources are booked by the minute. For you, this means: the minute the load peak has passed and you reduce the capacities, you will no longer be paying for them.

Server management using an API interface

Managing your virtual data center does not require use of the Data Center Designers (DCD). Thanks to our Cloud-REST-API, you can access the entire configuration even without the graphical user interface. For instance, you can write scripts to automate your scaling, e.g. throttle the system performance at night to release unneeded capacities. An effective way to save money! In our DevOps community, you will find a large number of microservices that let you automate your infrastructure just the way you need it.

More about the Compute Engine