Prometheus
Banned
- Joined
- Mar 19, 2006
- Messages
- 4,252
- Reaction score
- 3
Reason dictates they should. During power failures I've seen my phone change tower after a short while. Even in bad weather they sometimes seem to shut down so I was actually beginning to wonder if Vodacom doesn't use solar power as of late.Don't the towers have some form of UPS/backup generators? Even in quite extensive power failures I've still always had cell reception.
236.8k is definitely edge and not gprs. Something must be wrong. Remember that it shows the speed you connected at and does not update to the speed you are currently on.Just checked mine - shows 236.8k on GPRS. Either way - it is the port speed, and yes, no way they modem knows what the tower stuff is doing.
Incidently - have pushed MTN and Virgin sims in mine - says EDGE for them - which is correct.
I thought it was something like that. Nice post explaining it all.JimM, most of the sites nowdays have battery capacity of over 200Ah. With the bigger sites drawing some 50Amps plus from the 50V battery, you can see that it is not likely that the site will last more than a few hours if the power goes off. The rectifiers in most sites are capable of supplying over 60amps at 50V to run equipment/charge batteries when the power is on.
Then there is the other issue that the equipment can run for short periods without aircon, but after a couple of hours, it generally starts shutting down as the temperature rises to around 50deg C. Most sites are fed with a minimum of a 40A 3phase supply and some of the larger require a 60A supply.
Now what size UPS would one need to run an 18000 btu aircon and supply the power to run the equipment and charge batteries?
If only life was so easy.![]()
Yes but how much would direct power cost them and I suppose power isn't available everywhere like here. Generally power generated with an efficient generator wouldn't cost that much more, remember that all power originates from generators. However the initial investment in VERY expensive (this is not your ordinary household generators) generators isn't worth it for occasional non-essential backup power. I'm with them on this one, the money is much better spent on improving the network than using billions catering for a few sites going down occasionally.The numbers - on battery I assume no Aircon - then 2500-3000watt.
Total supply (40A one) includes running equipment, charging batteries and running aircon - appr 22-25kw depending on power factor assumption.
I am sure they can do better - MTN uses 10m litres of diesel in Nigeria a month - most of their base stations run off generators. And they are making money.
Surely they channel money spent on other cr@p to improve this issue.
What topic is this thread again??
Without power - no base station - no throughtput. guess still vaguely related![]()
The problem is not the latency as mentioned but the increase in latency (more than double) for larger packets that do not fragment. No it's not simple networking 101. The latency with the store and forward devices is already there, it's the extra latency which is the problem. Even at a clock of 100MHz and I can't see it being anything less than that a store and forward device operating on a 2Mbps link and in bytes no less can not have that much of an impact on the overall latency. If this was the case 10Gbps links would be non-existent. Your first post also said you wanted to know if larger packets had higher latency meaning it shouldn't necessarily be the case. It seems like the systems simply choke up on larger packets.I've before asked you to buy a book on networking and IP, and again implore you to do so. This is simple networking-101.
Fortunately, networks do run according to the laws of physics and thus can be understood by anyone willing to take the time to learn how it all works. You don't need to guess. It's all in the maths.
Latency is a function of transmission physics (~.7c/clock) plus the elements in the path (especially store and forward device, of which there are many where you must add the clock-in, processing and clock-out latencies, again dictated by physics) overlayed with a higher level protocol (such as TCP) that requires turn-around at some point (called handshaking) which resets the one-direction latency.
If you do the maths you'll come up with the latency you measured with something like a ping.
It's really not black magic and no-one is introducing latency just to be spiteful.
I can type up a (very long) post explaining it all, but it's just so well documented, anyone can read up on it. Google is your friend......