Scalable Cloud Computing Hosting with GoGrid

by Sergey Bushik and Katherine VasilegaAugust 3, 2010

This blog post explores GoGrid, a cloud computing hosting service, its capabilities, challenges associated with memory usage, backups, etc., and recommended use.

Table of Contents

GoGrid overview

The GoGrid name stands for the cloud computing hosting service for deploying and scaling server infrastructure on demand. It hosts Linux and Windows virtual machines managed by a multi-server control panel, supporting Windows Server, Red Hat, and CentOS Linux.

Pros

GoGrid is a user-friendly control panel with multiple options to create and manage servers. The panel features simplicity and high rate of speed for commissioning new hardware and setting up your servers. GoGrid enables a user to deploy a cloud server easily: just select a server image and a server size, and the instance will be up and running in minutes.
The platform enables flexible server image selection (CentOS, Red Hat, Windows) with some out-of-the-box support for Ruby on Rails and LAMP. GoGrid server images use standard Windows or Linux Operating Systems with full root/administrator access, giving the freedom to configure cloud server according to customer’s demands. Additionally, the system has an ability to deploy custom server images with MyGSI (My GoGrid Server Image). You can customize any of GoGrid maintained base images with own configuration and software and save it to personal Cloud Storage allotment. With the GoGrid Partner Server Images, some preconfigured software applications (solutions for development and testing, disaster recovery and backup, cloud management, security, monitoring and reporting) can be deployed quickly and easily with little or no configuration.
Continue with the advantages of fine-grained access to deployed GoGrid instances, including such aspects of control as role-based access, sub administrators, web-based control panel, remote access API, and managed DNS.
Among other GoGrid’s pricing benefits, its free inbound data transfer should be mentioned.
Other advantages provided by the solution include 24/7 support, 100% uptime service level agreement, and dedicated support teams.

Cons and recommended use

Still, the solution has some drawbacks associated with memory usage, backups, etc.:

One of the challenges is that a cloud provider ties the HD size (persistent storage) to the RAM size allocation, though base images for the lowest RAM size allocation becomes greater: for 512 MB is 30 GBs now. To reduce this challenge, the cloud provider announced its commercial cloud storage and a customer can provision cloud storage allotment from a user interface.
There are also constraints about load balancers per account allowed. Any attempts to add more than three load balancers will fail.
Furthermore, there’s a limited set of load balancer types supported by the cloud provider: Round Robin and Least Connect (the Fastest Response Time method). In fact, if server instances are similar in hardware specifications, the Perceptive, which predicts the server based on historical and current data, instead of the Least Connect method can be the best to use. If server resources are different, more flexible, the Weighted modification of the Round Robin method, which assigns requests to servers in turn according to their weights, may be a better solution.

GoGrid can be a suitable platform for hosting e-commerce, Web 2.0, and web applications throughout the development cycle—from initial development, to a staging environment for QA, to a full production environment.