OK. So I think we can agree that a SwOS box needs much more control over the switch chip than ROS has (assuming there’s a switch chip on the board), or it needs to bridge as required (and, if so, must be performant enough to do so). Given that the use of a switch chip in current RB products is a way of off-loading tasks which require no intelligence and that a SwOS box would need that intelligence, my money’s on there being no switch chip.
I haven’t had much luck with the on-board switch chip. I’ve had two different RB450G stop “switching” after a power failure, but the configuration was correct. I changed back to bridging and haven’t had any problems since.
I am guessing it will be a basic switch with a few extra bells and whistles compared to the current implementation on the RB4/7xx series.
The switching market is vicious, and very difficult to gain market share in. Over the years I have worked closely with Allied Telesis, Cisco, Foundry, Extreme, Juniper, HP ProCurve, Linksys and Nortel and all have had their ups and downs. Most try to differentiate themselves on features such as ESRP, MPLS, or into the different markets. e.g. Allied Telesis are big in schools and 3rd tier telco’s.
This product is potentially more interesting than even the RB1100 and I will be watching this one closely.
As far as I am concerned, the question is what features will SwOS have that ROS doesn’t have? The way I see it, I can’t think of a single one (but maybe I need to think harder).
So if we assume that SwOS is a cut-back version of ROS and further assume that the S250G comes in the same package as a RB750G - Given that a fully featured RB750G retails for about GBP50, a SwOS equivalent would need to be priced at around GBP30 or less I would have thought. That’s not a lot of money and not a lot of room for margin.
Assuming that the S250G follows the same naming convention as the RBs, it would have 5 ports. Is there a market for 5 port managed switches?
On the other hand, it could use a new numbering convention and actually be a 25 port device. One could postulate that if it was, port 25 would be SFP. That’d be nice.
Also it will be difficult for MikroTik to match e.g. HP in
30 years guarantee
Next business-day replacement
So they will have to go for
Price
Features
Given that a basic 24 port Gigabit HP switch costs 480€ (2510-24G) a MikroTik with MSTP,LACP,etc. and the bridge filter we have now, costing less would be an improvement
I think you will be right about MPLS, MikroTik is pushing it.
And if MikroTik does not do the “we sell you a SFP module worth 40€ for 400€ just beacuse we check the vendor-id of the module that is nevertheless produced by Finisar”-game I’m nearly sold.
And if MikroTik does not do the “we sell you a SFP module worth 40€ for 400€ just beacuse we check the vendor-id of the module that is nevertheless produced by Finisar”-game I’m nearly sold.
Lol, yep my thought exactly. A lot of that stuff if OEM’d out of Chinese OEM’s like Fiberay anyway.
Oh, and I thought I would mention:
A 24 port 10/100 edge switch should have at least 2 gigabit uplinks
A 48 port 10/100 edge switch should have at least 4 gigabit uplinks
A 24 port gigabit edge switch should have at least 2 10 gigabit uplinks
HP at least get that right for the most part.
Anyway, enough speculation lets see what Mikrotik announce!
The old RB naming convention is broken for switches. I want to see the number of ports at the end of the model number. Maybe they are teasing us and it will be S250G-12 or S250G-24. I really hope it is not just another 5 port device. If it is, it needs to sell for US $29 or less.
After the MUM announcement you will know everything about the RB250G. At that point, we will start a competition. Send in your suggestions and feature requests to support, and the best and most reasonable suggestion will receive a prize at the MUM. the winner must be present to receive the prize. The prize will be a RouterBOARD device.