OVERVIEW

The Need for Scalable File Serving Solutions

Explosion of Digital Content
Two major trends in the enterprise data center are driving the need for highly scalable file serving solutions. First, there is an explosion of digital content as information increasingly is stored and delivered in the form of digital images, audio & video files, PDFs, simulation and modeling data, email and other office documents, scanned images, and other types of "unstructured content". This new unstructured content varies in file structure and size. Traditional enterprise applications such as an ATM transaction typically create "structured" files that are a just a few kilobytes in size whereas unstructured content can range from a few kilobytes to even terabytes. More importantly, this content is being utilized as business critical data by most enterprises today. Overall, in excess of 5000 petabytes of disk storage shipped into the enterprise in 2006 alone as the consumption of storage capacity tracks this macro trend.

Move to Distributed Computing
The tremendous advancements in processing power of standards-based computing systems has fueled movement away from traditional monolithic, SMP-based computing architectures towards computing systems deployed in a "scale-out" or distributed manner. In addition, the enormous cost savings with this approach has allowed information-driven businesses to deploy computing infrastructures able to process significantly more information than ever before. Unfortunately, there is a snag preventing these businesses from utilizing these infrastructures efficiently for I/O delivery. Enterprise applications that once ran efficiently on a closed system (e.g. main frame) are starved for data in the new paradigm because traditional file serving solutions are woefully inadequate. This is further compounded by the ever increasing amount of unstructured digital content that is processed and shared by these distributed computing engines.

The I/O Delivery Challenge

Performance Scaling
The I/O delivery challenge has two facets:scalability of performance across mixed workloads and file sizes and scalability of capacity. Scaling on performance involves either scaling on bandwidth (MB/sec), .i.e. sequential access of large files using large I/O sizes or scaling on throughput (operations per second), .i.e. random or sequential access of small and medium sized files using small I/O sizes and involving lots of "metadata" operations such as checking the attributes of a file. Most scalable file serving solutions are designed purely for scaling on bandwidth; however, most enterprises have a variety of applications, some which require scaling on bandwidth, while others. that require scaling on throughput.

Capacity Scaling
The capacity scaling problem has to do with the need to build large file systems (a.k.a. namespaces) to store the increasing amounts of information being processed by legacy applications that rely on the information being stored in a single namespace or, in some cases, even a single directory of a file system. While many recent file serving solutions have the ability to address this amount of capacity in a single file system/namespace, it is most often not advisable to do so in practice due to the fact that these file systems are "all or nothing" i.e., a fault in a small area of the file system can render the whole file system inoperable for an extended period of time (hours, days).

Scale Performance and Capacity Independently
Finally, enterprises want the ability to scale performance and capacity independent of each other, provision them online without any disruption to their applications and yet provide the highest levels of availability for their applications.

Achieve significant increases in application performance,
Simplify management and operations of large data repositories,
Implement a pay-as-you grow business model without forklift upgrades, and
Maintain the highest levels of availability and resiliency for their mission critical applications.

IBRIX Fusion

IBRIX Fusion scalable file serving solutions are specifically designed to meet the challenges of today's demanding compute environments. These solutions enable customers to:

IBRIX Fusion gives you the ability to dramatically improve the performance of your mission critical applications. Our customers are leveraging IBRIX Fusion to shorten product development times, increase productivity, perform faster data analysis and make more accurate business decisions in industries as diverse as media and entertainment, financial services, automotive/aviation and biotechnology just to name a few. The IBRIX Fusion scalable file serving solution is equally optimized for scaling on bandwidth and throughput and can thus scale sequential bandwidth from 100s of MB/sec to 100s of GB/sec and scale throughput from 10s of thousands of I/Os per second to Millions of IOPS.

Customers are also leveraging IBRIX Fusion's ability to scale capacity to create low cost bulk storage solutions that enable them to start small with a few terabytes but simply and easily scale to Petabytes of data all while providing comprehensive single namespace management, yet with the highest resiliency, fault-tolerance and fault-containment.

IBRIX Fusion also allows enterprises to scale performance and capacity independently and non-disruptively, thus providing a flexible pay-as-you-grow model that scales with the growth of your business.
IBRIX Fusion scalable file serving solutions are also highly available providing monitoring and fault isolation and allowing for failover at the component and system level. Performance and capacity can be provisioned online and non-disruptively without affecting the uptime and SLA for your applications.

Infrastructure Solutions