Swarm Tools on different Infrastructure

[ashishmundra4001]

If I have a Strands Agent (lets name it WebCrawler) which can browse the web pages and extract data. Let's say this Agent is deployed in AgentCore Runtime Instance A.

Now I create a new Strands Agent (lets call it Coordinator) that wants to browse through 100 web pages. So, it uses Swarm Tool and invokes WebCrawler Agent.

Does Strands Swarm SDK has capability to then dynamically create lets say 10 AgentCore Runtime Instances A1, A2, A3, .... A10 and give each instance of Agent 10 unique urls which they can process in parallel and return the Summaries to this Coordinator Agent?

Or does Swarm here means that Coordinator will just have one Instance of WebCrawler Agent i.e. Instance A which will run 10 instances of WebBrowserTool on instance A of AgentCore Runtime and do the work parallelly on same Compute - A?

I am trying to understand that how is this Swarm Intelligence Tool different than SemanticKernel equivalent Orchestration in terms of using AgentCore capabilities?