WAN optimization

from Wikipedia, the free encyclopedia

The term WAN optimization describes technical mechanisms that contribute to the smooth transport of data packets over a wide area network , or WAN for short.

Purpose of WAN optimization

The optimization technologies are necessary because a large number of data of different types, sizes and priorities in the context of applications (applications) pass the wide area networks. In addition to e-mail traffic, business-critical, bidirectional applications such as ERP or database systems also run through wide area networks. Voice applications such as IP telephony (Voice over IP) and video conference systems, all of which run in real time , are also operated via the networks. All of these applications must be available with a certain performance in order to run smoothly at the individual locations of large organizations. Since the bandwidth in wide area networks is usually limited, the individual applications can compete with one another, which could impair their performance and thus jeopardize business continuity. WAN optimization technologies counteract this.

Development of WAN optimization solutions

Functions for WAN optimization are bundled in product solutions from various manufacturers. The solutions mostly consist of software and hardware components. Here, at the individual sites of an organization (at the interface from the WAN to the local networks ( Local Area Network (BMS short) installed shortly LAN)) bandwidth management systems. They are used to measure, control and optimize network traffic based on the defined performance goals. Central management software controls the WAN optimization process that encompasses all locations. The software usually also provides the performance targets and configures the decentralized BMS. In addition, it collects all measurement data recorded by the BMS and provides a real-time interface for access to this measurement data.

WAN optimization with CoS / QoS

WAN optimization solutions use different mechanisms to channel the data packets in the WAN. When a technology is the existing bandwidth to multiple power classes ( Classes of Service , in short CoS) with a certain quality of service ( Quality of Service , in short: QoS) distributed. However, this technology is only suitable to a limited extent for WAN optimization.

Strategic approach

A modern generation of WAN optimization systems, on the other hand, is based on the strategic business goals of an organization and combines functions for transparency, optimization and acceleration of network traffic in a holistic approach. The focus is on the precise control of the network and the applications, the prioritization of business-critical applications, the dynamic allocation of the optimal bandwidth and the acceleration of all data packets. The solutions are also auto-adaptive, which means that they always adapt flexibly to the respective conditions in the network.

Features of modern WAN optimization solutions

transparency

The network managers of large companies with several geographically distributed locations must be able to understand at any time which applications are running with which performance over the WAN. Modern WAN optimization solutions therefore provide maximum transparency about current network events. Monitoring functions measure the current performance of individual traffic flows in detail and in real time. Each data packet is analyzed in the entire network traffic up to the application layer (layer 7 of the OSI reference model ). The network managers receive information about delays (delay), runtime differences ( jitter ), packet losses and the throughput of each individual data stream. Only in this way can agreed Service Level Agreements (SLAs) for the individual company applications be documented, permanently checked and ensured.

Auto-adaptive optimization of network resources

The latest generation of WAN optimization solutions take into account the company's business goals, the requirements of the applications and the desired user experience . You therefore forego the rigid classification of MPLS technology and instead use intelligent mechanisms for dynamic allocation of bandwidth. Using certain criteria, usage classes are defined for the individual applications with corresponding internal SLAs in order to determine the respective network-wide performance targets and priorities. On the basis of measured real-time data, the solutions dynamically compare the current bandwidth requirement with the currently available network capacity, calculate the rules for the currently required optimization every second, taking into account the specified performance targets, and carry them out automatically in real time at each location. Compared with the fixed bandwidth allocation of conventional optimization methods, this adaptive approach enables optimal use of the available network capacity at all times, even with heavily meshed traffic. This helps to ensure that users are always guaranteed the performance levels defined in advance and that the agreed QoS are achieved.

acceleration

To ensure optimal application performance, preferred data packets should be routed through the WAN as quickly as possible. Because especially with transactional and real-time applications, minimal response times are of crucial importance for perfect functioning. A SAP or Oracle system that works too slowly, for example, has a considerable negative impact on employee productivity. In real-time applications such as Voice over IP or video conferences, time delays can severely disrupt communication. However, the acceleration of a certain application traffic must not be at the expense of the quality of other, equally important data streams. A complete, automated monitoring of the network traffic and thus also of the various acceleration mechanisms is essential. The permanent measurement of the data streams means that it is always known how much bandwidth is available and how much is required. Traffic is also precisely analyzed in order to differentiate between business-critical, less important or pure leisure applications. The WAN optimization systems then decide independently, depending on the criticality, which applications are accelerated through the network, taking into account the available bandwidth.

technology

To compensate for the delay (latency) on WAN lines, there are various approaches that are used individually or in combination on the WAN optimizers (WAN appliances):

  • TCP / IP: Various TCP / IP protocol limitations are circumvented by optimizing the transmission characteristics, for example loss handling, window scale, congestion notification, etc.
  • Intermediate storage / compression: data packets can be identified using hash values . In this way, content that has already been transferred can be temporarily stored on the appliance (similar to a proxy ) and only the significantly smaller hash value has to be transferred.
  • Protocol optimization: some application protocols are designed for LAN speed and are accordingly chatty . This may not be a problem in the LAN environment, but in the WAN environment many small packets are always occupied with the latency of the WAN line. To counteract this effect, individual application protocols are optimized on the appliances so that the chatty communication is reduced to a few TCP packets.

Provider of WAN optimization systems

Important providers in the international market for WAN optimization systems are, for example:

Open source solutions

  • OpenNOP
  • WAN proxy
  • TrafficSqueezer

Web links

WAN Acceleration Technologies : The Fundamentals of Wide Area Networking