Today’s enterprise networks are facing unprecedented demands – streaming videos, employee-owned mobile devices and collaboration tools are all contributing to far more traffic than email, FTP and general Web surfing produced just a few years ago. How much? Consider that business IP traffic is projected to increase at a compound annual growth rate of 21% from 2012 to 2017 – that’s faster than IP WAN (13%), but only a fraction of the projected growth rate for mobile business Internet (59%).
What issues does this create for enterprise network administrators? Increased usage leads to networks that are congested and can’t scale, and frustrated end users who just want their applications to perform faster and better.
Administrators responsible for addressing network congestion often resort to one of three approaches: adding bandwidth, accelerating network traffic, or cutting off access to applications believed to be causing congestion.
Adding bandwidth can address the issue, but it’s a short-term fix that creates two problems. First, because network usage continues to increase, IT budgets also continue to increase in perpetuity, which isn’t sustainable. Second, increases in bandwidth demand have outpaced decreases in bandwidth cost.
Just adding bandwidth treats all applications the same – from workplace productivity tools to visits to YouTube. In effect, network administrators are funding the increases in recreational traffic, which only temporarily and partially solves the problem.
Another common approach is application acceleration. Some organisations invest in solutions that accelerate everything on the network, including unwanted and unproductive recreational traffic. Just like adding more bandwidth, this approach treats all traffic the same. In a sense, organisations are spending money to improve not only the speed of business-critical applications, but that of recreational applications as well.
Finally, cutting off access for applications believed to be causing congestion is not feasible either. Sure, people are watching cat videos on YouTube, which you will want to control. However, the CEO could be utilising video for a demo on the fiscal results call. Treating all users and use cases the same is a partial, inelegant solution at best.
To reduce network congestion without funding inefficient solutions or imposing heavy-handed restrictions on end users, organisations need to manage network traffic at a segment level. Doing so enables network managers to create policies that optimise and protect the performance of business-critical applications while limiting the impact of recreational applications. By taking this approach, network managers can reclaim bandwidth for business-critical applications and defer costly upgrades.
The following are seven ways you can reduce network congestion by managing network traffic at a granular level:
#1. Look into the network: What’s the problem? Who’s causing it?
The first step in solving congestion problems is determining what’s causing congestion. Find out which applications are in use, how much traffic is generated by each, and the relative priorities of these applications. Identify whether there are high-traffic periods during the day or month and whether some applications have special quality-of-service (QoS) requirements. Without an accurate understanding of current traffic flows, any action to relieve congestion will be only a guess.
#2. Reduce traffic via caching
Cache frequently accessed Web objects locally so only new ones are downloaded. If multiple requests are made for the same objects multiple times, the overall bandwidth utilisation accumulates. For instance, patch bundles or applications can be downloaded from an external site once, and subsequent accesses will benefit from proximity. Whether performed serially or in parallel, every hit adds up and can affect performance or volume-based fees. Using caching strategically can improve performance and lower utilisation. A combination of browser caching, LAN caching and perimeter (proxy-based) caching will reduce wasteful repetition of data transfers.
#3. Control recreational traffic
Place limits on recreational traffic during work hours that will still allow access to recreational sites but grant priority to business applications. Even if a company’s Web policy allows personal and recreational use, or if incidental use is overlooked, non-business content can become an unneeded burden on the network. Through traffic shaping, QoS rules, utilisation caps or blocking, availability can be assured for business critical needs.
#4. Time-shift your network
Are certain applications regularly utilised at defined times? Create policies that reflect this, and shift resources to those applications. Analysis of the applications in use and the nature of the traffic they generate can uncover groups that compete for scarce resources and indicate whether they would benefit from simple scheduling changes. For instance, backup windows could be adjusted to prevent interfering with batch data transfers or replication transports. This reduction in network contention can significantly improve end-to-end performance and usability.
#5. Don’t treat all business traffic the same
Categorise traffic into three buckets:
* Reputational: Items that would have a direct impact on current or future business, based on customer perceptions about availability and performance.
* Revenue-affecting: Items that would prevent or decrease the amount or rate of income, whether based on sales or production/fulfilment.
* Mission-critical internal: Items that are necessary to support functions not purely customer facing or profit generating but that are needed for daily operations and ongoing business functions.
Reputational traffic should be the initial focus, but not to the detriment of the remaining traffic.
#6. Manage and prioritise at a user level
Determine who the user is and what application is being used. Is the CEO using videoconferencing to speak to key investors? Where is the user? Is he logged in remotely?
#7. After investigating current packet flows, create an intelligent policy
Allocate network resources based on a predetermined set of criteria. For instance, during the hour that the CEO is making a videoconferencing call, 40% of the organisation’s available bandwidth might be allocated to the call, while 15% is allocated to a marketing video and the other 45% is dedicated to other activities. The policy will generate different priorities on other days, such as near the end of the quarter, when accounting applications may receive the highest priority.
Understanding current traffic patterns and establishing priorities requires a powerful management tool. It must monitor and report packet flows and sources of delay, and provide controls to set and enforce priorities. Despite the need for this functionality, only 35% of IT managers have invested in such a tool while 81% have increased bandwidth.
Exinda WAN Optimization Suite
Exinda’s WAN Optimization Suite, with its Network Control Suite component, provides a unified view of the network. Links, applications and users are all visible through a single Web-based management screen.
The Exinda solution continuously monitors the network, ensuring that each user and application is allocated the network resources needed to meet productivity goals. It integrates with Microsoft Active Directory to report usage by individual user and group and display both total usage and usage by application.
Exinda’s Application SLA Assurance enables network managers to create policies for users and departments, roles and applications, locations, times, devices and network segments. Reports make clear how well the network is meeting end-user needs. During high-traffic periods, critical applications are granted the resources needed, and policies define the maximum bandwidth allocated to each application.
Exinda’s solution makes cloud migration easier with displays that monitor and report cloud provider service-level agreements. As applications are moved to the cloud, the tool monitors performance, ensuring that end users continue to receive the level of performance promised by the provider.
Application performance score technology helps locate problems by providing a detailed view of network performance. For each application, score technology measures TCP packet round-trip time; network, server and transaction delay; the number of packets lost; and network jitter. Performance scores are continuously monitored and displayed in real time. Summary reports show performance scores by user, application, transaction and server. If performance deteriorates, the reports make clear whether delays are introduced by network bottlenecks, slow server performance or some other source.
WAN performance is critical
The move to converged networks and cloud applications has made WAN performance a critical factor. Branch office users must experience quick response from Microsoft Exchange servers in a central office. Cloud-based Software-as-a-Service applications must also provide consistent rapid response.
Exinda’s WAN acceleration technology has been proven to reduce WAN traffic by 90% or more. It compresses data and eliminates redundant WAN transfers. As widely dispersed groups update shared files, only the changes are moved across the WAN. Repeated Web access is avoided by caching commonly used pages at locations throughout the network. Only the modified portions of Web pages are downloaded for each access.
Slow application response affects mobile users, just as it does central office or branch employees. Exinda’s Mobile Workforce Solution reduces traffic by applying the same caching and compression techniques as the WAN solution. Application-specific modules optimise savings using techniques suited to the selected application.
Recreational traffic has proved to be a common source of WAN congestion. WAN Optimization Suite’s interface with Microsoft Active Directory makes it apparent who is misusing the network. URL tracking separates legitimate Web access from Web cruising. Deep packet inspection identifies evasive applications, which are often used to cloak access to file-sharing sites.
Anonymous Proxies Detection prevents users from circumventing URL tracking. Detection is challenging because new anonymous proxies appear every day. Exinda’s appliances automatically download proxy updates each day from Exinda to maintain up-to-date detection. Exinda products are available as standalone or virtual appliances. Models range in capacity from units suitable for a small branch office up to a large data centre. Appliances can be clustered to provide high availability. The company’s tools provide the same management capabilities for IPv6 traffic as for IPv4, so when your network transitions to IPv6, the same Exinda tools will continue to provide the same detailed reporting and control.
In the past, network managers have lacked visibility into actual network performance. End users reported poor application performance, but hard-to-use network tools made it difficult to determine the source of the backups. Until now, network managers have had no choice but to add bandwidth, a solution that adds expense but often does not address the problems. With Exinda WAN Optimization Suite, it’s possible to pinpoint the source of the delays and implement network policies that deliver excellent performance.
For more information contact Exinda South Africa, pieter.olivier@exinda.com, www.exinda.com
Source: http://go.exinda.com/7-tips-to-reduce-network-congestion.html. Used with permission.
© Technews Publishing (Pty) Ltd. | All Rights Reserved.
printer friendly version