I have gone through the instructions about 4 times in setting up a MediaCache server and I'm still either not understanding the concept or configuring something wrong. I have been following the "Scale video on demand (VOD) with MediaCache" document. I currently have 24 Wowza streaming engines deployed internally and they received requests via F5 Global Traffic Manager boxes so the requesting users is routed to their local server. We run about 3 live events per month internally with around 1,000 live viewers. I want to know leverage those 24 boxes to do some VOD replays in a similar manner.
Do I have the proper understanding of how the content flow should work?
- I'm assuming that I can put an MP4 on a LAMP box and point all the VOD Edge configurations for the media cache to that server and they will pull the content as needed. So instead of 500 users making requests of one server. I want 24 servers to make those request and then cache that content for a couple of days.
MC1 <-> User
MC2 <-> User
HTTP Server <-> MC3 <-> User
MC4 <-> User
MC5 <-> User
If this is correct then I'm not sure what I'm doing wrong. I have run sniffers on the HTTP server looking for requests from the MC servers but it never receives any. I'm a bit at a loss right now. But I wanted to make sure I understood the concept properly before I go in depth on the configuration.
MediaCache basically fetches VOD data from its defined source, which can via HTTP fetches from a web server or FILE (e.g. NFS/SMB) from a file server. Putting your MP4s in a valid HTTP path would work fine. The Wowza Vod Edge servers maintain a local store of fetched data so any subsequent request is from there rather than the defined source.
It should be quite straightforward to set up an HTTP source (e.g. Prefix: http/ Base URL:
http://mywebserver/) which would result in the stream name http/sample.mp4 being used by MediaCache as
http://mywebserver/sample.mp4 (if it isn't already cached).