alex_aldo - Fotolia
There's a fine line between under- and overprovisioning network resources when designing enterprise networks. Yet, for many network architects, the fear of not having enough network provisioning is greater than the fear of having too much.
This trepidation can lead to networks with far more performance and throughput capacity than required. This little-discussed practice of network overprovisioning wastes funds that could be better spent elsewhere.
Because networks commonly last between three and seven years -- and because upgrades can be complex, disruptive and expensive -- network professionals should buy and deploy slightly more networking resources than required throughout the network's lifecycle.
Network provisioning and capacity planning
Designing an enterprise network requires a bit of research, input from key business leaders and following a few steps to avoid overprovisioning.
Network capacity planning tips include the following four steps:
- Review network baselines.
- Review the IT roadmap.
- Understand network hardware limits.
- Determine a business pivot.
1. Review network baselines
Organizations planning to upgrade their existing networks should start the network capacity planning process. Network professionals should review network monitoring tools using legacy Simple Network Management Protocol (SNMP) or modern network analytics platforms like AI for IT operations (AIOps).
These tools monitor network component CPU, memory and usage levels and map network activity over time. They contain a wealth of historical network usage and performance data on which to base network capacity decisions.
If historical data is available, for months or even years, network architects can calculate the average increase in network usage and performance over time. Network teams can use this information to forecast future performance requirements of the network.
Network professionals should also review network bandwidth, especially at the internet and WAN edge, with historical information to measure whether throughput performance could reach a point where it impedes business operations.
Network professionals should also use this time to shop around for alternative internet and WAN connectivity options. Emerging technologies, like 5G, could be available at a reasonable price point, and in many cases, a multiyear contract for connectivity services comes with discounted pricing. It's preferable to calculate the amount of throughput needed for today, as well as for the months and years ahead.
2. Review the IT roadmap
While historical data is useful, teams should also factor in any future technology projects and their effect on network performance demands. Common IT projects -- such as cloud migrations, IoT deployments and big data analytics rollouts -- can cause massive changes in network throughput demand that can increase, decrease or change the flow of data throughout the corporate network.
For example, if the IT roadmap shows that apps and data currently managed on premises will eventually move to a public cloud, then client and server data flows will change. These flows will no longer route to the on-premises data center. Rather, they'll travel through the public internet or direct WAN links to public cloud services.
This type of roadmap change will dramatically shift performance and throughput needs on campus routers and switches. This means that, as the IT infrastructure architecture changes, the location of where teams provision network resources may also change.
3. Understand network hardware limits
One way to understand hardware and software limitations, including maximum network throughput, is to reference the network vendor's data sheets that outline all capabilities, constraints and upgrade possibilities.
For example, chassis-based network switches often come with the ability to upgrade their core processors over time to provide increased backplane throughput. However, the chassis itself has an absolute maximum throughput capacity it can handle, regardless of any processor upgrades. Network professionals should understand the boundaries to make the most cost-effective decisions when upgrades are necessary.
Along those same lines, switch stacks have become a popular option in access and distribution layers of an enterprise network. With a stack, however, backplane capacity is permanent and cannot be upgraded, unlike chassis-based switch hardware. Network architects should keep in mind that, when it comes time to add additional switch hardware to the stack, each added switch will consume more throughput that will eventually reach stack backplane limitations.
4. Determine a business pivot
The fourth, and often the most difficult, component to review when right-sizing a network is to gauge the stability of the business. Network professionals should consider the chance their businesses may need to pivot network management strategies. Ideally, the business strategy is well established, which means the network demands aren't subject to change. In this situation, teams shouldn't allocate much overprovisioning.
On the other hand, if the business strategy is prone to change, this creates a provisioning risk in the network capacity planning process. A pivot in strategy can require more -- or fewer -- networking resources. Thus, business leaders need to determine how much overprovisioning risk they want to take. They need to anticipate unforeseen demand versus the risk of wasting resources if the change in business strategy never happens.
This is where what-if scenarios should play out. Organizations need to determine the likelihood of change and the cost-benefit analysis of each scenario as it relates to the network.
Weigh the cost of provisioning and capacity planning
Prior to starting this process, network teams should anticipate the cost of researching and planning for future network provisioning and capacity. While inherent knowledge of business needs and future outcomes is an important part of network provisioning and capacity planning, network visibility tools make the job easier and less time-consuming.
Of course, network visibility tools also cost money and time to purchase and install. The good news, though, is the cost required to set up monitoring and analytics tools, like SNMP, AIOps and NetFlow analyzers, is worth the investment. In addition to monitoring, these tools can troubleshoot network outages and performance issues. Most organizations are willing to implement and use these tools for a host of purposes, including network provisioning and capacity planning.