 |
|||||||
|
Wide Area Network Optimization As technology evolves the workforce continues to demand that networks keep up with increasingly varied needs. Greater speed and connectivity, virtualization of the work environment, flexibility and interoperability are not just special features anymore but a requirement, a must-have in today’s business environment. Rapidly changing business factors are driving the need to manage and optimize bandwidth. Users have grown to expect a consistent level of service regardless of their proximity to the resources (servers, storage, etc.) they are connected to. Mobile devices, work at home programs and virtual employees have extended the traditional building- and campus-wide networks into large scale global networks with rapidly changing traffic patterns, applications and protocols (collaboration, voice and video). To remain competitive businesses must be flexible through improved services while being mindful of capital expenditures. For this adequate tools are necessary to ensure quality and performance is not suffering due to increasing demands for content delivery and Web 2.0 rich media technology. Today’s networks support various types of applications with various protocols each with varying degrees of sensitivity to jitter, delay and packet loss. Organizations have rapidly adopted content delivery technologies such as SharePoint and web user interfaces to deliver applications to their users and customers through common web browsers. This places an increasing demand on WAN performance, requiring administrators to manage their available bandwidth. Prioritization and acceleration techniques offer the ability to improve performance, reduce bandwidth utilization, deliver quality of service and most importantly reduce costly bandwidth and circuit upgrades. Bandwidth optimization is not a new concept. Traditionally, vendors have offered Quality of Service (application prioritization - Packeter’s packet shaper) and bandwidth compression software and hardware as point products. These products focused on a specific aspect of WAN optimization typically addressing only one aspect of network performance with limited or no awareness of the application layer. Today’s offerings have combined these series of technologies coupling traditional methods such as:
New approaches to solving the WAN bandwidth problem combine these traditional methods with new techniques such as Local Instance Networking (LIN) (also known as bit caching). Local Instance Networking inspects all WAN traffic, storing a local instance of the data bit pattern within a hierarchical tree as an independent data store/ cached at each enterprise location. The local instance is populated transparently to the user and is updated in real time. This allows each data set to be stored only once per location, permitting weeks of data to be stored. When a repeat of a previously transmitted bit pattern is queued for transmission the local acceleration appliance suppresses the transmission, replacing it with the index number of the data store location at the remote end. This reduces the amount of bits transmitted and ensures repetitive information is not sent across the WAN. At the file level, LIN enables only the changes to be transmitted across the WAN instead of the entire file, maximizing bandwidth utilization and application performance while protecting the integrity of the data. Additionally, many vendors have incorporated add on features to improve the performance of unique protocols. These features include support for HTTP, TCP and CIFS. These are particularly important when addressing large amounts of chatty or highly transactional protocols. Many users are mobile and are no longer confined to corporate offices. These traveling and mobile users require access to information resources from public and private networks with non traditional wireless devices such as cell phones, PDAs and blackberry devices. To address this need, some vendors offer client technology as an agent loaded on the individual’s computer. This agent extends WAN optimization technology beyond traditional boundaries and improves application response times. When deployed properly, these appliances and agents deliver the performance of distributed servers, without the cost or complexity and at a demonstrated acceleration rate of up to 80x traditional transport rates. Although these product offerings show great promise, they are not a one size fits all technology. Not all vendors offer the same method or advance protocol support. It is important to understand the types of data and the protocols in use on their networks prior to evaluating / selecting vendor technologies. The current environment does not support the use of different vendors at different sites, so in selecting a vendor product one must match in order to maximize the performance gain based on the specific profile of their network. The importance of the availability of a software client agent to meet their specific requirements should also be considered. Current bandwidth requirements need to be taken into account when analyzing the performance of specific appliances. Some vendors offer better performance than others, alleviating the need to stack multiple appliances in a load balanced format to obtain desired throughput performance. The placement of these devices must also be carefully considered. They must be positioned on the clear side of any encryption appliance. For these devices, consideration should be given to performance and any potential security issues they may present. It is highly recommended that a full network assessment be performed before implementing a solution so as to rule out any possible issues, encryption, frame size mismatch, etc.
|
USAGE | PRIVACY POLICY