By Amichai Lesser
Networking is getting tougher. Networks must deliver a growingrange of services, from ERP, CRM and email to VoIP and webservices applications, each of which has its own idiosyncrasiesand requirements. Each new service introduced onto the networkcontends for available resources with every other service,impacting the network’s ability to support the business.
Meanwhile, the network itself is constantly changing. Newlocations are added – some of which may be in another country oron another continent. Equipment is upgraded and/orre-configured. New management and/or security tools maythemselves impact service performance. Decisions about datacenter consolidation and business re-organization also affectthe network in different ways. All of this makes the network ahighly dynamic environment where even subtle changes can have amajor, unforeseen impact on application performance andavailability.
Yet business users expect this complex environment to be asreliable as electricity – despite the fact that networkingbudgets are not being increased in proportion to these growingchallenges. So network managers can’t simply over-provisionnetwork infrastructure to make sure every service has all thebandwidth it needs. Moreover, over provisioning may not evensolve the problem and/or ensure the required level ofperformance.
That’s why network managers are facing many challenges,including:
1) Pinpointing potential networkperformance issues early in the development lifecycleIdeally, the impact of the network on a new application orservice should be dealt with from the very beginning of thedevelopment process – when potential problems are much easierand less expensive to fix. Unfortunately, this is rarely thecase. Problems with an application’s “networkability” aretypically discovered only after its roll-out into the productionenvironment is initiated. At that point, it’s usually too lateto make any significant changes in the application’s design. Sothe problem gets pushed onto the shoulders of the networkingteam. That’s why, in ’05, smart network managers will focus onnipping these problems in the bud.
2) Validating new or modified applications and infrastructurebefore they are deployed in production As the network becomesmore complex and more critical to the day-to-day-operation ofthe business, networkperformance related risks associated with application andinfrastructure change are continuing to rise. In fact, some ofthe worst business interruptions that companies havehistorically experienced have not been the result of unexpectedequipment failure. They’ve been the unexpected consequence of aplanned modification. Networking teams must therefore implementchange management best practices in ’05 that prevent them fromhaving to put out fires that they accidentally startedthemselves.
3) Improved troubleshooting of intermittent/transient networkproblems One of the most frustrating things for a networkmanager is dealing with a problem that keeps disappearing beforeit can be adequately understood and remedied. However, as thebusiness’s tolerance for network interruptions continues todrop, these intermittent problems will become a biggermanagement issue. So this year, network management teams need todevelop more effective methods for capturing transient networkconditions and discovering the root causes of these problems.
4) Accelerated time-to-benefit for new and/or upgradedapplications When C-level executives decide to make investmentsin new applications and services, they want to see thoseinvestments pay off quickly. That’s why the slow, stagedproduction roll-outs of the past won’t cut it anymore. Instead,networking teams need to be able to quickly deploy newapplications across the enterprise. This can only happen ifcaution and uncertainty about the actual behavior of theseapplications in the production environment is replaced byconfidence and certainty in ’05.
5) More intelligent planning for and support of business growthNetwork managers constantly have to cope with change. They haveto determine how increases in network utilization will affectapplication performance. They have to decide how to bestengineer the network to support business expansion,re-organization or mergers and acquisitions. However, they canonly do so if they have an effective means of performingcapacity planning tasks and assessing a full range of “what-if”scenarios. Such scenarios are also critical for formulatingrealistic contingency plans that can ensure business continuityunder a variety of possible conditions.
Looking at these challenges, it quickly becomes evident thatconventional production network management tools alone are nolonger sufficient for today’s networking teams. These tools aregreat for monitoring the production network and discoveringcertain types of problems – but they don’t enable networkmanagers to validate new technologies and applications beforethey’re deployed on the production network. They also forcenetwork managers to solve problems that should have beenaddressed in application design.
Conventional tools aren’t very helpful for troubleshootingintermittent and/or transient network problems either, sincethey don’t provide a means of reconstructing and analyzing suchintermittent conditions. Nor do they help accelerate productionroll-outs, facilitate experimentation with “what-if” scenarios,or support formulation of network contingency plans.
So what’s an overworked, under-resourced network manager to do?The answer is to look at network modeling technologies. Thesetechnologies provide an environment in which new applications,technologies and problem-solving strategies can be safely andthoroughly evaluated. Because they allow an application’snetwork behavior to be fully validated before it’s deployed inthe production environment, these technologies also empowernetwork managers to perform more rapid, glitch-free roll-outs.Plus, modeling technologies are uniquely able to provide insightinto any number of “what-if” scenarios – so network managers canmake plans for growth, corporate re-structuring and/or disasterrecovery.
“Empirical” modeling solutions offer today’s network managementteams particularly excellent business value, because of theiraccuracy and relative ease of implementation. This accuracy andease is achieved by running the actual applications against amodel that uses captured conditions from the productionenvironment. The result is a clear understanding of the userexperience well ahead of deployment.
For this and other network performance articles, white papers,and industry resources, please visit Shunra athttp://www.shunra.com/resource_center.aspx.
About Shunra
Shunra’s solutions empower organizations to address servicelevel and performance concerns before rollout. The Shunra VEsolution creates an exact replica of the production networkenvironment, enabling IT professionals to safely develop, testand experiment with applications and infrastructure beforedeployment, and effectively plan for growth and change. Tailoredfor networking, performance and testing professionals, andsoftware developers, Shunra VE facilitates collaboration acrossIT disciplines so IT organizations can quickly and moreefficiently uncover and resolve problems before they impact thebusiness. Over 1,500 leading enterprises and technology vendorsworldwide are using Shunra’s award-winning solutions including3M, Boeing, Cisco, Dow Chemical, EMC, FedEx, General Electric,General Motors, JPMorgan Chase, Kelly Services, Merrill Lynch,Motorola, Nestle, Pitney Bowes, and Vodafone. Shunra’sheadquarters are located in New York City and Kfar Saba, Israel,with worldwide offices in the UK, Sweden and India. Shunra isalso supported through a global network of channel partners.
Article Source: www.ArticlesBase.com
