С. Перроте (Q Logic) Ускорение работы приложений

Editor's Notes

1. Today I’d like to provide you with an overall update on QLogic, our product and technology direction and our overall state of business operations. I’d like to make this a very interactive, collaborative and open discussion, so please feel free to ask questions throughout the course of this presentation.
2. Higher bandwidth networks and high performance servers with increasing core densities are driving more server virtualization. Performance hungry business critical applications like OTLP, ERP, Analytics, Mail server, web 2.0, VDI are driving more demand on storage I/O. Increased server performance (core density, functional integration) = Increasing “IO Intensity” = Demanding more transactions and less waiting, want Lower latency and response time: more data = more revenue (SLAs!)…effects the bottom line Disks are getting bigger, not faster, driving a large I/O performance gap, impacting the application response time, no improvement in TPS. In recent years, with emergence of affordable and reliable SSD technology, Storage I/O Performance gap is being addressed. SSDs provide very high performance for a small, random IO workloads, driven by the enterprise applications. Thus, SSD technology can be an effective solution to the I/O performance gap problem. Enterprise applications are clustered and virtualized Faster application response time results in increased productivity, and improved customer experience SSD technology can play a pivotal role in solving the IO performance gap problem which is inhibiting the applications ability to deliver faster responses. Let’s look at how different solutions using SSD technology are evolving to solve this problem……….
3. Considering the need for shared Data model, early solutions used SSDs in the Storage array controller. Simple to manage and implement. Data remained shared, available and protected. IO Performance improved BUT…. Multiple network hops between server and storage array controller and bottlenecks in the storage array controllers themselves could not realize the true SSD performance. Therefore, the next step in evolution was to……….
4. ……Deploy SSDs in the server and use them as cache. It brought critical Data close to the application, performance benefits are fully realized!!!!!!. BUT…… More SW added to the server to manage the cache tier, it became more complex. Captive cache in a server broke the “shared data” data model. Even with great performance benefit this step remains incomplete as it breaks the “shared data model”. We said, Enterprise applications require shared Data model and high performance. Let’s look at what is required to keep the benefits of SAN and gain the benefits of SSDs while managing the TCO.
5. Read slide title Accelerated Caching for a Server Cache Shared Data Lun Multi Server…..
6. Thank you for your time today and thank you for your interest in QLogic and our broad portfolio of networking solutions.