HyperCache

Verivue's Intelligent Cluster Technology is a high performance caching file system that fragments and distributes objects across a cluster of commodity servers using state-of-the-art consistent hashing algorithms. Distribution of content across the cluster prevents hot spots and automatically handles flash crowds, regardless of differences in customer library size or popularity. Other key features and benefits include:

Support for a variety of workloads

Our HyperCache solution supports a variety of HTTP workloads, providing whole site delivery from a single platform with no performance penalty. With whole site delivery, there is no need to deploy a separate caching infrastructure for each type of content. A single cache will support a range of applications from ad serving, e-commerce and photo sharing to audio/video streaming and software downloads.

Whole site delivery provides efficient delivery of various file sizes and types, including:

• Small Objects - HTML, thumbnail images, JavaScript objects, style sheets, text files and generally any object under 1MB
• Large Objects - audio and video files delivered via Progressive Download (PDL), software downloads and generally any object over 1MB
• Video - HTTP adaptive streaming across all formats, including Microsoft's Smooth Streaming and Apple's HTTP Live Streaming.

Unmatched Scalability

The HyperCache scale out architecture is a best-of-breed design approach for web-scale systems. Unlike single node systems offered by most competitors, bandwidth and storage capacity can be expanded in virtually limitless fashion by adding and interconnecting individual nodes to form a "cluster". All nodes in the cluster are connected via Ethernet and have full read/write access to storage across all other nodes, thereby forming a common store.

Network wide scalability can be achieved by establishing parent-child relationships across clusters arranged in a hierarchy. With this model, caches can resolve misses either directly through the origin or through other caches in the network.

High Availability

An advanced self healing design ensures continuous availability of the cluster even after a part of the system fails. All remaining components seamlessly subsume the workload of the failed component, preventing hotspots and providing non-stop operation. Server clusters can be deployed in places where on-site support may not be regularly available.

Self Load Balancing/Self Organizing

HyperCache features a self load-balancing mechanism which evenly distributes workload across all nodes in the cluster. This feature helps eliminate performance hotspots without the need for external load balancers, saving cost and complexity while reducing energy consumption and rack space.

As new nodes are added to the cluster, the HyperCache automatically redistributes objects, seamlessly integrating the additional bandwidth and storage capacity into its overall resources. Capacity planning is simplified, since no explicit mapping of content to nodes is required.

Low-Latency Design ideal for live streaming

HyperCache employs a unique cut-through routing method that virtually eliminates delay in the handling of cache misses, a critical requirement for live HTTP streaming applications. Live video can be pulled through several levels of cache hierarchy with minimum latency. In contrast, store and forward techniques used by most competitors can introduce several seconds of delay at each level. The feature eliminates the need to pre-position content on remote origin servers, simplifying planning and ongoing maintenance.

Flexibility of choice

While most competitors bundle proprietary server appliances with their application software, Verivue offers a software-based solution that harnesses the advantages of industry-standard server and storage components, offering a more flexible alternative to the fixed appliance model.

As a Hewlett Packard authorized reseller, Verivue offers numerous choices in platforms and cluster configurations. Clusters can be configured to meet a wide variety of needs and accommodate an array of workloads with a broad range of server platforms (rack mounted or blade servers), storage solutions and networking/communications interfaces.