With our current special to nearly double the RAM included with our standard nodes (and newly created RAM nodes) we decided to talk a little more about RAM. This post is one of three posts dedicated to our friend in the cloud; RAM.
Virtual Private Servers (VPS), or cloud servers, radically distinguish themselves from both shared servers and dedicated servers in allowing administrators the ability to quickly scale CPU, RAM and storage resources up or down, at a moments notice, providing an almost unparalleled level of control, versatility and cost savings in the Internet era.
While all three computational elements mentioned above are essential to the health and performance of your server, RAM is arguably the most critical (and flexible) resource to quickly boost the speed and performance of a website or application.
RAM stands for random access memory and in a nutshell is the workspace on the server. Information stored in RAM/memory (don't separate the two or you'll get destroyed by your geek friends) will be returned quicker to the user or application than that stored on the hard drive. Thus, the more space for RAM storage, the less often the server needs to access the hard drive.
While there are additional technical challenges to consider when increasing RAM, including Auto Scaling on Linux-based environments, this route is considered one of the easiest and most practical ways to supercharge performance on cloud servers.
Generally, speaking current VPS packages will allow you to select an entry-level cloud server which allocates an initial amount of RAM (e.g. 512 MB) but can be scaled up – or down – at any point. This can be a godsend when a web server experiences a seasonal spike in activity during the Christmas shopping period or the launch of a new e-commerce application that finds itself under unexpected pressures from more buyers purchasing products through the system.
Most cloud applications require at least 1GB of RAM to run efficiently but can be tuned to run on less. In other cases, high-volume servers may need between 2 to 4GB of RAM. Additional RAM will allow more files to be cached thus reducing disk requests.
The type of operating system and web server selected on a cloud server will also influence these decisions. For instance, Windows administrators running IIS installations should ensure the operating system paging file size matches the amount of RAM on the server.
Today, most cloud server admins will also ensure their customers are covered if a sudden spike in activity (temporarily) pushes their website or application past the reserved RAM initially purchased in the system. This spike allocation is referred to as burstable RAM and a represents a huge improvement over the memory restrictions encompassing a shared server or even a dedicated server. The latter server is handicapped because administrators generally need to overprovision the box to factor in future spikes in activity at the time of purchase. This can make them inefficient and costly in relation to the up-and-down-RAM-scaling available in cloud servers.
Due to the rapid and exponential growth in web application usage at both the video and gaming level, spread across a range of devices, including smartphones and tablets, many developers and ISVs are choosing to utilize a cloud server to give them the flexibility, business agility and cost-savings required to scale their RAM resources up or down as market demand fluctuates throughout the course of a calendar year.