ObjectStor Cloud Storage

Introduction

NephoScale’s ObjectStor cloud storage, otherwise known as file based cloud storage, is a highly scalable and redundant cloud storage service that is designed to compliment our Cloudlet virtual servers and BareMetal Dedicated Cloud Servers to make complex computing applications easier to implement and scale. The NephoScale Object-Based Cloud Storage service is a web based storage system which provides a cost disruptive, ultra reliable, highly redundant storage infrastructure designed to support mission critical storage applications using a simple RESTful web services API to store and retrieve objects.

Overview

With NephoScale ObjectStor cloud storage, you have the ability to create multiple slices (storage partitions) and multiple keys (user authentication) to support many different users within one account. You can also manage these cloud slices & keys, view storage metrics, and billing metrics all through the NephoScale portal. If you want to set-up several cloud storage slices at one time, you can do this by making a single API call using our CloudScript. ObjecStor supports bindings in many popular languages (e.g. PHP, Python, and Java) that can easily interact with our cloud storage API, allowing developers to work with our cloud storage solution immediately. By authenticating with the proper credentials, you can store (write) and retrieve (read) files on NephoScale’s cloud storage infrastructure from anywhere in the world. Like most enterprise NAS systems, it can be used as mountable storage (familiar to most administrators) similar to a NAS like mounted directory, allowing all existing NAS applications to utilize the service without any modification. ObjectStor cloud storage is perfect for serving static content for web applications, storage of large public or private data sets, such as online files and media, and archival and back-up storage.

Advantages

An advantage that the NephoScale Object-Based Storage solution has over enterprise storage solutions is the ability to scale to hundreds of petabytes and beyond. Traditional file systems, like NAS architectures, involve a hierarchical structure that can limit their overall scalability and negatively impact performance as they grow. Traditional NAS systems reach a tipping point before getting to even one petabyte of data.  Our object based cloud storage infrastructure has been built with a modular approach, using auto-scaling load balancers, proxies, and storage arrays. When our system detects an increase in data store, an increase in web requests, or an increase in bandwidth requirements, it can scale out any of the modular elements accordingly. This means the system can scale virtually limitless without significant degradation to performance.

Another advantage that our cloud storage solution has over an enterprise solution is redundancy. We currently replicate every object three times across different object stores (physical storage arrays) in different zones (physical locations in the datacenter). We do checksum analysis to detect data corruption and automatically repair it using redundant data. If one of the objects becomes corrupt, the system can determine the consistency of all three files, remove the corrupt file, and restore that object using the other two. This translates to uptimes in excess of 99.999%.