Question:
What are the bandwidth requirements to run cloudscape? How will this impact our network?
Answer:
For Discovery and Inventory phase specifically, the best practice is to start with a smaller subnet range and expand the scope with iterative scans. Not due to performance impact on the network, but due to having erroneous or out of scope systems clutter the Asset Inventory report.
Size of communication from the RN150 and the servers scales depending on how busy the servers are. For instance, a server with thousands of connections will give us a large `netstat` output. On RISC test environment, between 27 and 81 MB communication was observed every 24 hours for a single server. Each collection sweep during this test was less than 1 MB for a fairly idle server, and 3 MB for a much more busy one.
The bandwidth will depend on how much data is moved. Each major data set is transferred separately, but it is compressed and encrypted prior to sending. Assume less than ~100MB per transfer.
If network stability is at risk or of a major concern, please start small in the scope of scanning. This can be done by using a single subnet section for inventory as is mentioned above and having the onsite team monitor the increase in network or server load during that time. Those same devices could then be licensed and monitored for 2-3 days to examine load during performance collection.
On engagements of similar size, we noticed users performed a full collection of the IT estate at the same time and did so without any issue. In larger environments of 10+ thousand, we saw a significant drop in this practices, and users started with a smaller 500-1000 server scan/performance collection first for 2-3 days, then rolled out collection to the remaining servers.
