www serwers optymalization
Architecture
Crescendo’s Maestro Platform is based on a unique purpose-built architecture which represents the industry’s first implementation of Layer 2-7 functionality in dedicated hardware. Designed from the ground up to handle complex acceleration/optimization tasks, the architecture uses multiple network processors along with multiple FPGAs (Field Programmable Gate Arrays) to accomplish all processing tasks. Within the main processing components, 80 Application Micro Engines (AMEs) are employed to perform multiple functions available in the device. Each AME is responsible for a different task and all AMEs operate simultaneously. The result is a truly distributed architecture where the complex processing tasks are spread across multiple components, each with dedicated resources. Other architectural highlights include:
- All functions, including TCP termination/offload/acceleration, load balancing, compression, and SSL offload are performed by dedicated hardware
- Dedicated memory is allocated to each component so there is no memory shared between tasks
- A totally hardware-based compression engine performs inline compression at multi-gigabit scale, with zero latency. The Use of programmable components allows new features to be added via software upgrades, rather than component replacement
- The platform is designed to provide maximum scale, while remaining flexible. This creates a robust, hardware-based platform with hardware performance and software flexibility.
Maestro's distributed, hardware-based platform not only delivers maximum speed but provides an architecture that allows features to be enabled concurrently without any performance penalties. This Feature Concurrency is unattainable with standard appliance designs that use a central processing/memory pool for all functions.
NetWorkWorld Reprints
One of the major benefits of Maestro’s distributed hardware architecture is the concept of “feature concurrency.” Feature concurrency is the ability to enable multiple functions/features without losing performance. Today’s acceleration/optimization products are seeing more and more functions consolidated into a single device. What often goes unnoticed is the effect enabling multiple functions has on the system as a whole.
Traditional appliances are based on a shared-resource architecture where off-the-shelf components are used together with a standard operating system (and possibly some offload hardware). Functions that are implemented on such a model all share the same resources (i.e. processing pool, memory, etc.). With each feature that is enabled, more and more of these shared resources are used, almost linearly affecting the overall performance of the device as a whole. The system (and therefore, the application it’s front-ending) suffer in performance, then, when multiple functions are operating simultaneously. This is a vital fact that is often ignored when discussing acceleration/optimization appliances.
Maestro’s unique purpose built hardware architecture uses dedicated components (each with dedicated resources) for each of its functions — so no single feature affects the performance of any other feature. This means that enabling multiple functions has no detrimental effect on the performance of the platform, allowing the application to use all the benefits of the Maestro functionality with no performance penalty. This is one reason why Crescendo’s Maestro platforms offer performance orders of magnitude above competing solutions.
|
The CN-5504 offers the Maestro platform's award winning hardware architecture and performance in a four port unit, with an optional redundant power supply. The CN-5504 offers the ... |
|
|
|
