| Networking is getting tougher. Networks must deliver | | | | and discovering the root causes of these problems. |
| a growing range of services, from ERP, CRM and | | | | 4) Accelerated time-to-benefit for new and/or |
| email to VoIP and web services applications, each of | | | | upgraded applications |
| which has its own idiosyncrasies and requirements. | | | | When C-level executives decide to make |
| Each new service introduced onto the network | | | | investments in new applications and services, they |
| contends for available resources with every other | | | | want to see those investments pay off quickly. |
| service, impacting the network's ability to support the | | | | That's why the slow, staged production roll-outs of |
| business. | | | | the past won't cut it anymore. Instead, networking |
| Meanwhile, the network itself is constantly changing. | | | | teams need to be able to quickly deploy new |
| New locations are added - some of which may be in | | | | applications across the enterprise. This can only |
| another country or on another continent. Equipment | | | | happen if caution and uncertainty about the actual |
| is upgraded and/or re-configured. New management | | | | behavior of these applications in the production |
| and/or security tools may themselves impact service | | | | environment is replaced by confidence and certainty |
| performance. Decisions about data center | | | | in '05. |
| consolidation and business re-organization also affect | | | | 5) More intelligent planning for and support of |
| the network in different ways. All of this makes the | | | | business growth |
| network a highly dynamic environment where even | | | | Network managers constantly have to cope with |
| subtle changes can have a major, unforeseen impact | | | | change. They have to determine how increases in |
| on application performance and availability. | | | | network utilization will affect application performance. |
| Yet business users expect this complex environment | | | | They have to decide how to best engineer the |
| to be as reliable as electricity - despite the fact that | | | | network to support business expansion, |
| networking budgets are not being increased in | | | | re-organization or mergers and acquisitions. However, |
| proportion to these growing challenges. So network | | | | they can only do so if they have an effective means |
| managers can't simply over-provision network | | | | of performing capacity planning tasks and assessing a |
| infrastructure to make sure every service has all the | | | | full range of "what-if" scenarios. Such scenarios are |
| bandwidth it needs. Moreover, over provisioning may | | | | also critical for formulating realistic contingency plans |
| not even solve the problem and/or ensure the | | | | that can ensure business continuity under a variety |
| required level of performance. | | | | of possible conditions. |
| That's why network managers are facing many | | | | Looking at these challenges, it quickly becomes |
| challenges, including: | | | | evident that conventional production network |
| 1) Pinpointing potential network performance issues | | | | management tools alone are no longer sufficient for |
| early in the development lifecycle | | | | today's networking teams. These tools are great for |
| Ideally, the impact of the network on a new | | | | monitoring the production network and discovering |
| application or service should be dealt with from the | | | | certain types of problems - but they don't enable |
| very beginning of the development process - when | | | | network managers to validate new technologies and |
| potential problems are much easier and less | | | | applications before they're deployed on the |
| expensive to fix. Unfortunately, this is rarely the case. | | | | production network. They also force network |
| Problems with an application's "networkability" are | | | | managers to solve problems that should have been |
| typically discovered only after its roll-out into the | | | | addressed in application design. |
| production environment is initiated. At that point, it's | | | | Conventional tools aren't very helpful for |
| usually too late to make any significant changes in | | | | troubleshooting intermittent and/or transient network |
| the application's design. So the problem gets pushed | | | | problems either, since they don't provide a means of |
| onto the shoulders of the networking team. That's | | | | reconstructing and analyzing such intermittent |
| why, in '05, smart network managers will focus on | | | | conditions. Nor do they help accelerate production |
| nipping these problems in the bud. | | | | roll-outs, facilitate experimentation with "what-if" |
| 2) Validating new or modified applications and | | | | scenarios, or support formulation of network |
| infrastructure before they are deployed in production | | | | contingency plans. |
| As the network becomes more complex and more | | | | So what's an overworked, under-resourced network |
| critical to the day-to-day-operation of the business, | | | | manager to do? The answer is to look at network |
| network performance related risks associated with | | | | modeling technologies. These technologies provide an |
| application and infrastructure change are continuing to | | | | environment in which new applications, technologies |
| rise. In fact, some of the worst business interruptions | | | | and problem-solving strategies can be safely and |
| that companies have historically experienced have | | | | thoroughly evaluated. Because they allow an |
| not been the result of unexpected equipment failure. | | | | application's network behavior to be fully validated |
| They've been the unexpected consequence of a | | | | before it's deployed in the production environment, |
| planned modification. Networking teams must | | | | these technologies also empower network managers |
| therefore implement change management best | | | | to perform more rapid, glitch-free roll-outs. Plus, |
| practices in '05 that prevent them from having to put | | | | modeling technologies are uniquely able to provide |
| out fires that they accidentally started themselves. | | | | insight into any number of "what-if" scenarios - so |
| 3) Improved troubleshooting of intermittent/transient | | | | network managers can make plans for growth, |
| network problems | | | | corporate re-structuring and/or disaster recovery. |
| One of the most frustrating things for a network | | | | "Empirical" modeling solutions offer today's network |
| manager is dealing with a problem that keeps | | | | management teams particularly excellent business |
| disappearing before it can be adequately understood | | | | value, because of their accuracy and relative ease of |
| and remedied. However, as the business's tolerance | | | | implementation. This accuracy and ease is achieved |
| for network interruptions continues to drop, these | | | | by running the actual applications against a model that |
| intermittent problems will become a bigger | | | | uses captured conditions from the production |
| management issue. So this year, network | | | | environment. The result is a clear understanding of |
| management teams need to develop more effective | | | | the user experience well ahead of deployment. |
| methods for capturing transient network conditions | | | | |