in panagenda news, Trainings

If you’ve ever found yourself tapping your fingers, waiting for a database to open, you know that network bandwidth is an important part of IT. What may not be obvious is the calculation behind exactly how much bandwidth is needed to meet end-user demand, nor how to calculate this value for a consolidation project.
While Service Level Agreements (SLA) or operational guidelines may exist for server and even application operation, knowledge around bandwidth sizing for an IBM Domino consolidation or centralization project rarely does.

Such project must start with a detailed analysis of the existing environment, in order to calculate and map the current demand into the new server and network configuration as the most important element of a successful consolidation is the network bandwidth between the new datacenter and the remaining user locations.

With our clients, we often have discussions where organizations simply calculate future demand, by taking the amount of traffic they see with network monitoring tools on network switches. The flaw with this is that those switches don’t cover the real end user traffic because that often remains on the LAN level, which is behind those network switches. In essence, when dealing with an IBM Domino server consolidation, information coming from network monitoring solutions cannot be used for any calculation.

So how do you accurately analyze the demand pattern while distinguishing between user and server network consumption?
Turns out network bandwidth consumption between servers is obsolete if the aim is to centralize all servers, as pictured in the below. This leaves user bandwidth consumption as THE key information you need to look for.

2_2-Measuring-Workload_5

Currently, panagenda is the only company in the World which can provide this level of detail for your consolidation project. Even better, our analysis reaches beyond bandwidth consumption levels and into port compression and distinction between heavy and average users, user activity on databases and more, which means we’re able to propose different network link scenarios for our customers.

The following matrix provides a sample calculation for network bandwidth required for each.

Matrix of Sample calculation for network bandwidth

The used capacity columns show network demand coming from end users. Splitting user demand up by heavy and average users provides the result shown on the right, where used capacity is compared with current bandwidth which tells you if you can or cannot centralize server from City X towards a central location.

Addressing heavy user impact on peak and network demand is particularly important because they have a significant impact on your future environment, as shown in the following picture:

2_3-Measuring-Workload_2

Understanding peaks, who is causing them and where those users are coming from is key to predicting how many servers can be centralized and which locations are best suited for such consolidation. And as illustrated in the next graphic, when considering shared bandwidth, which is not 100% dedicated for Notes and Domino, a further calculation is needed to understand consolidation implications and potential.

20-08-2014 11-14-28

While planning a consolidation and ending up with an efficient end result requires more than these network considerations, it’s a great start towards ending up with an environment that meets end user demands, while being more cost effective to run.

Want to learn more about network bandwidth or other consolidation topics?
Drop us a line here or attend one of our webinars!

Related Posts


Leave a Comment

Start typing and press Enter to search

iDNA Demo Report