Model library
Prepared models that open in the simulator, ready to run. Nothing to install and no account — press play and watch, or run a batch for figures with confidence intervals. The same models are behind the Model Library button in the simulator's toolbar. More are on the way.
One server, one line
Arrivals at 0.8 a second against a server that handles 1 a second — busy 80% of the time. The waiting line grows and drains, and nothing is ever turned away.
Try this: Raise the arrival rate towards the service rate and watch the waiting time stop rising in a straight line.
Open in the simulatorThree servers, one line
Twice the arrivals, shared by three servers that each handle 1 a second — the same 67% busy, but one queue instead of three.
Try this: Compare it with three separate lines. One shared line is far shorter, and almost nobody predicts by how much.
Open in the simulatorA line that fills up
Room for five waiting and no more. An arrival that finds it full is turned away rather than queued, so the system loses work instead of growing a backlog.
Try this: Push the arrival rate up and watch what happens to the throughput, not just to the queue.
Open in the simulatorThree workers, one database
Arrivals at 4 a second into one line, three workers at 2 a second each, all handing over to a Shared Server named Database that slows the more it holds — about 0.56 seconds a request here, against 0.33 alone.
Try this: Raise the Database's Contention to 0.3 and its Coherency to 0.05 and watch the requests inside it pile up without limit.
Open in the simulatorTwo paths, one pool
Orders at 8 a second and reports at 1 a second share a Resource Pool of 5 connections. An order needs one for 0.2 seconds, a report for 3 — so an order waits about 1.5 seconds for a connection, and between 2 and 3 in every 100 requests give up after 5.
Try this: Raise the pool's Capacity to 8 and watch the line in front of it drain, although fewer than 5 connections are in use on average.
Open in the simulatorA surge of orders drains the pool
The same pool of 5 connections, with orders on a traffic profile: 4 a second for a minute, a climb to 20 a second held for 30 seconds, then back to 4. Reports stay at 1 a second throughout, and both paths wait on the same connections.
Try this: Switch on the orders' Arrivals Against Profile chart, then watch the line for a connection build through the surge and how long it takes to clear once the orders fall back.
Open in the simulatorChange a model and your version is kept in the browser like any other work — in place of the model you had there before, so save that to a file first if you want it back. Open a model's address again and it is back as published. Or start from a blank canvas in the simulator.