You can soon rent a LiveU box that can stream HD over cellular networks. It has HD/SD-SDI in and Analog and Firewire inputs. Just patch your mixer outputs to that box and you can stream without using ethernet or Wi-Fi. (even has optional satellite connectivity)

http://www.liveu.tv/hd_series.htm

I see your point about the flexibility between PC versus dedicated streaming hardware. One thing I might point out though is that the Niagara 4100 is full updatable via Firmware download. They figured out the problems they could run into a while back. so they did what everyone else did which is allow keeping things current by downloads or firmware and patches from the web.

Hey I'm a lover of big in PC's also. I've got a rack several Dell PowerEdge servers all have tri-network connectivity. They talking over 8Gb/sec FC, 10GbE and of course conventional 1GbE. I have a 8Gb/sec FC SAN on the network that gives me access to lots of fast FC disk. I even got into fast FC connected LTO-3 tape libraries to help archive the HD stuff away. I also can do real time dual stream uncompressed HD SDI capture (Black Magic product) if required.

So no argument from me about PC's. Long live the Dell Workstations with Dual Quads :-)

Yes you were right I also have Server & PC redundency. I think the advantage of the Niagara is similar to Run&Gun scenarios with my news gathering cams. kind of like Run&Stream capability :-)

Glad to hear that - Viewcast is a quality company, I've always been happy with their products. I developed prejudices against this sort of hardware solution early in my career, it just always hurt like he** to be looking at $20k worth of AVID-certified-and-supplied hardware wondering how to get rid of it because the new stuff was gonna' require different AVID-certified-and-supplied hardware. I don't mean to pick on AVID, they did what they had to, as did many very successful companies.

However, the cost and, um, immediate flexibility arguments remain, at least in my market as a boutique webcaster. Another example, one of these days, a client is going to require HTML5 in a live webcast; when that day comes I'll have a choice of several software encoders. If I were to depend on a hardware encoder, the firmware for HTML5 may or may not exist at that point, it might not be configurable to client requirement.

The example of the broadcaster is a good one - they're not going to go HTML5 until everything is worked out, the firmware has been through a couple releases, etc.

You know, that's a really good point. You can add an encoder appliance to an existing team, and not have to add another body to make it work. The approach that works for me is different - I get to be a (mostly) dedicated webcaster, with my fingers on the various buttons. Staffing has to be considered, and is frequently the largest concern.