As the need for Web 2.0 and grid computing environments becomes mainstream, the demands for more cost-efficient, scalable and flexible file storage solutions has intensified. File system capacity requirements for storage is soaring into the petabyte range and the I/O demands on storage in performance compute environments are approaching billions of IOPs/sec, and managing such massive amounts of data is putting a strain on IT resources. Although the scalability needs are different in today’s bulk (low cost, high density) and performance (high bandwidth and high throughput) file storage environments, the file system deployed to meet the challenges of these environments doesn’t have to be different.
The seminar will discuss a segmented, parallel software architecture approach that is breaking the barriers in capacity, performance, management and cost of conventional file solutions. Attendees will learn the advantages and disadvantages of conventional file serving approaches and gain insight into how significant improvements in data delivery from the storage environment to computing elements, and tremendous flexibility in scaling capacity and performance independently, are achievable. Customer case studies presented will describe the advantages of a segmented, parallel file-system approach and walk through the approaches used in building an infrastructure that will improve:
- Performance: 10 to 50 times application performance improvements
- Capacity: scalability to multiple petabytes in a single namespace
- Management: dramatically lower the cost of storing and managing large volumes of data, allowing smaller, disparate file systems to be consolidated into a few easy-to manage namespaces
- Cost: hardware independent architecture that will attach to commodity or best of breed storage hardware, achieving the best price, performance and density requirements, allowing flexibility and choice, eliminating vendor lock-in while supporting the broadest set of industry standards