d
Size: a a a
2019 October 26
2019 October 27
SM
Вернемся опять на корветы/укнц/спики )
МС-1502 и ЕС-1840 😂
SM
А, Д3-28 ещё помню - из "персоналок"
АР
и перейти на русские мобилы
Н
Есть вичат
V
есть у кого APC типа SURT8000/10000 или аналогичные модели, включенные в трёхфазном режиме? Что они в интерфейсе показывают? Входное напряжение 220 или 380? Или они только первую фазу измеряют всегда?
Межфазную напругу вроде не показывает.. Только по каждом. Если не нравится 380 - переключается на первую фазу.. Поэтому если несколько девайсов стоит - надо разносить фазы
||
> a 10ge-4-port card can handle a maximum of 20 gigs.. but 67% of
> that comes to about 13.4 gigs
> since it is live traffic, the frame size of each packet is not
> fixed..
> but i took an average for the packet size as 1500
> the throughput that was got by the customer was 1.1Mpps.. which
> means, 1.1 * 1500 * 8 Mbps
> which comes to 13.2
> the values seem to be matching..
> so thus we confirm that for live traffic, when the frame size of
> each packet is not fixed, overal thorughput is about 67% from
> theoretical maximum of 20G
> using this as the base, we are able to found out there is
> 1.1Mpps limitation
>
> Support of 40Gig traffic on 10GE-4-port card depends on various
> traffic conditions.
> If we send pure L3 routed traffic with packet size of 1000
> bytes, then we can achieve, 10Gig ingress and 10Gig egress on
> each traffic port as we verified it in our lab.
>
> For bridged traffic, the results that can be achieved are a lot
> less. Ex: with 256 byte packets, we can achieve up to 72%
> throughput. But, with 64 byte packets the number is as low as 38%.
> If we look at the configuration used by customer (old – where
> only one 10ge-4-port card was used), there are two ports used
> for bridging (5/1 and 5/3 – interface Megafon-bridge). Hence,
> depending on packet size, a low throughput can be expected.
> The new configuration (Where customer has two 10ge-4-port
> cards), there is no bridging involved. All traffic is routed. In
> this case, a higher throughput is observed.
>
> that comes to about 13.4 gigs
> since it is live traffic, the frame size of each packet is not
> fixed..
> but i took an average for the packet size as 1500
> the throughput that was got by the customer was 1.1Mpps.. which
> means, 1.1 * 1500 * 8 Mbps
> which comes to 13.2
> the values seem to be matching..
> so thus we confirm that for live traffic, when the frame size of
> each packet is not fixed, overal thorughput is about 67% from
> theoretical maximum of 20G
> using this as the base, we are able to found out there is
> 1.1Mpps limitation
>
> Support of 40Gig traffic on 10GE-4-port card depends on various
> traffic conditions.
> If we send pure L3 routed traffic with packet size of 1000
> bytes, then we can achieve, 10Gig ingress and 10Gig egress on
> each traffic port as we verified it in our lab.
>
> For bridged traffic, the results that can be achieved are a lot
> less. Ex: with 256 byte packets, we can achieve up to 72%
> throughput. But, with 64 byte packets the number is as low as 38%.
> If we look at the configuration used by customer (old – where
> only one 10ge-4-port card was used), there are two ports used
> for bridging (5/1 and 5/3 – interface Megafon-bridge). Hence,
> depending on packet size, a low throughput can be expected.
> The new configuration (Where customer has two 10ge-4-port
> cards), there is no bridging involved. All traffic is routed. In
> this case, a higher throughput is observed.
>
Вот интересный топик, как раз на канал залез спросить у коллег кто какие методы тестирования каналов использует