Adaptive-Learning Bandwidth Optimization®
Adaptive bandwidth optimization for Makani starts with the premise of being application aware. By understanding the fundamental characteritics of specific apps like an email program or a web browser or an enterprise application, and by learning about the network link characteristics, you can make intelligent decisions on how to streamline an application for optimal delivery. Adaptive-Learning optimization starts by identifying the application and the network, understanding the inherent properties and the data model of that application, and then applying application and network-specific techniques to optimize link bandwidth. It is not just about traditional compression but an intelligent set of operations triggered by first identifying the type of the application and learning about the network.

The Makani Latency Buster™ Architecture
Makani appliances are intelligent in that they intelligently optimize and prioritize realtime and business-critical data for reducing latency. The Makani Latency Buster™ is a revolutionary architecture consisting of 5 key building blocks:

• Split Control-Plane®: Makani decouples control from the data path for IP-based transport protocols and apps. By separating control from data, Makani delivers a breakthrough system architecture that is highly optimized for control and data.

• File Transformers: Unlike generic compressors, File or Object Transformers correlate objects at the information level and employ patent-pending algorithms to remove redundancy that gets the best storage, bandwidth, and latency savings.

• Virtual Channels® Abstraction: A novel resource abstraction- Virtual Channels- is implemented to virtualize the link bandwidth. Virtual Channels enable fine-grained QoS control to offer the fastest acceleration for real-time and business critical data.

• Protocol Transducers: Makani optimizes and accelerates protocols at all levels. Besides TCP window resizing and other transport layer optimizations, a patent-pending Async Messaging Layer is employed that mitigates app and protocol "chatti-ness" and round-trips to deliver substantial reduction in latency.

• Hierarchical Memory®: By employing object- and byte- level data store with hierarchical hash technology, Makani detects data duplicates leading to significant bandwidth savings and manifold reductions in application latency.

MakOS™ (Makani Operating System)
Makani builds on the power of a customized Linux operating system. The MakOS™ architecture offers a "software bus & backplane"-style construct that is highly distributed and designed to provide high availability, flexibility, and performance. The MakOS™ achieves this by dividing up the processing into distributed modular, light-weight "processes" that separates network data processing from storage. By judiciously distributing load of the group of processes across all processors in the system, overall performance is significantly improved. The distributed design provides exceptional performance, ease of management, and service protection while fault containment significantly limits impact when a fault occurs.

Results are impressive. In Makani’s adaptive bandwidth optimization solution, initial bandwidth savings range from 40% to well over 90% depending on the type and mix of applications in-use. In conjunction with the revolutionary Makani's Latency Buster™ system architecture, it is not unreasonable to expect that almost any organization could optimize and save bandwidth and latency, to deliver 10X-100X data acceleration for bandwidths they already have, without changing the network infrastructure that are already in place.

Learn more about the Makani Global IntelliUpdates™ Network>>