L******t
发帖数: 1985 | 1 But you still got much wider bandwidth in datacenter. :)
In my case, I often see 20k download but 50k upload. Poor me. But that proves
that asymmetry doesn't play as big a role as bandwidth alone here.
Assume all others have opened up port forwarding when applicable. Is there any
difference for me to open up or not? That's what I'm wondering now. |
|
g**d
发帖数: 77 | 2 我对wireless懂的不多，要是说错了，不要在意。
The max throughput of wimax per cell site is just 70Mbps. Now it is
considered as 3rd pipe after teleco and cableco. But its physical limition
is a big issue to me. Now DSL/Cable could provide at 6mbps, the coming VZ's
FIOS is at 100mbps, and new DOCSIS 3.0 provides similar bandwidth. So wimax
is not a good fit for heavy bandwidth applicaiton like iptv or vod.
there is 2-3 articles on telephony margazine on Clearwire and Sprint's Wimax
projects. very interesting.
Sprin |
|
s*****g
发帖数: 1055 | 3 For IP FRR, this is true because IGP will converge no matter what, but for
MPLS facility-backup FRR, since the path is pre-determined in the ingress,
ingress may not implement re-optimization of protected LSP, i.e, protected
LSP may or may not be adaptive, so traffic may stay in bypass path,
bandwidth requirement is not guaranteed especially N LSPs are riding on the
same bypass.
For detour FRR, bandwidth constrain is not a problem by nature. |
|
s*****g
发帖数: 1055 | 4 For IP FRR, this is true because IGP will converge no matter what, but for
MPLS facility-backup FRR, since the path is pre-determined in the ingress,
ingress may not implement re-optimization of protected LSP, i.e, protected
LSP may or may not be adaptive, so traffic may stay in bypass path,
bandwidth requirement is not guaranteed especially N LSPs are riding on the
same bypass.
For detour FRR, bandwidth constrain is not a problem by nature. |
|
g******i
发帖数: 32 | 5 假设Internet有2个计算机，已知bandwidth等信息，如果一个计算机向另一个传送一个
任意大小的文件，如何估计传送的时间？
做这个计算除了bandwidth外，还需要知道什么已知条件么？简单起见，也可以假设是
基于TCP的。
如果能给出相关文献就更好了，俺自己去看：）多谢多谢 |
|
t*********e
发帖数: 1136 | 6 I'm not as well-informed as zher. But I think networking equipment industry
is in devolution, not evolution, despite some pockets of exciting
opportunities.
Long time ago, the networking industry faced many new challenges month-to-
month. There were a lot of excitement and a lot of demand. Nobody knew what
would be the final answer. The problems were also very big.
Right now no such big problems exist. Companies are fighting on little
things desperately in order to compete with each other.
Many ... 阅读全帖 |
|
s******v
发帖数: 4495 | 7 dish，cableco和teleco，其实是contents reseller，re-pkg and charge premium.
如果每家都是100M bandwidth, 其实最好是，直接去disney，fox，或者像netflix这样
的把内容放在一起。不过cableco/telcco不干，因为这样他们就沦落为utility公司就
像bandwidth，没有video pkg，margin就低了,但是他们也不敢说，block netflix,这
样美国FCC会介入，所以早晚的事情，ott会成为主流。 |
|
z**r
发帖数: 17771 | 8 2 things you are missing:
1. there are always multiple tcp sessions in the real world
2. without congestion avoidance like RED, whenever there is a drop, it's
tail drop.
So, when the link is congested, all the packets that are not in the queue
yet will be tail dropped, and these packets are from multiple tcp sessions.
Thus results in transmission wave or global sync, big waste of bandwidth (
queue actually)
With congestion avoidance like RED/WRED/FRED/etc., the lower priority
packets with higher... 阅读全帖 |
|
|
x*********n
发帖数: 28013 | 10 这个配置有什么问题吗？
interface GigabitEthernet0/0
description **** Fiber interface ****
bandwidth 150000
no ip address
ip pim sparse-dense-mode
load-interval 30
duplex auto
speed auto
media-type sfp
no negotiation auto
service-policy output QOS
!
interface GigabitEthernet0/0.70
description **** Fiber interface ****
bandwidth 150000
encapsulation dot1Q 70
ip address wan 255.255.255.248
no ip redirects
no ip unreachables
no ip proxy-arp
谢谢！
! |
|
a***o
发帖数: 3734 | 11 1. freenas根本用不了2G memory，偶的freenas只有512M，
拿出一部分来ramdisk装embedded os，剩下的跑emule，bt，
同时serve两部电影，还有50% free。
2. freenas是based on freebsd 6的，对有些sata controller
和网卡的支持有限，去
http://www.freenas.org/index.php?Itemid=30&id=freenas_users_hardware&option=com_openwiki
看看。
3.如果可以从usb启动就没必要花钱搞CF+CF-IDE adapter。一个usb key drive
足矣，即使买也不超过5刀。
4.p3的旧货足矣。craigslist上30刀，运气好还有free的。
5.一个pci-sata卡恶霸上只要7刀包shipping。
6.千兆网卡真的有必要吗？wireless的bandwidth和pci的bandwidth摆在那里呢。
7.机箱和硬盘散热很重要，毕竟你的nas最值钱的就是硬盘。 |
|
a***o
发帖数: 3734 | 12 展开说说？我一直觉的wireless的bandwidth就那么几十M，
emule/bt也不过几百K/s,ISP也就十几M的bandwidth，
gigabyte的网卡没什么必要。
换个角度想，4G一部电影，90分钟，平均50M per minute，
1M/s....
难道是偶太理想化了？ |
|
p**i
发帖数: 688 | 13 Please be aware slow speed is most likely result of your uplink bandwidth
since it typically is < 50KB/s for cable/DSL. If you have an active
uploading
task running without any limits, it will exhaust your uplink bandwidth for
sure. |
|
x*z
发帖数: 1010 | 14 你这个不会是bandwidth设错了吧，跟我老当年的错误一样
自己掐自己脖子玩儿？
先不要用任何QoS设置，找几个测速网站好好测测，搞清楚
真实的bandwidth，然后按那个设，另外你要设VoIP或者Skype
的话不知道你是怎么设的，反正不能按port，因为port都是
inbound不是outbound，而QoS只能设outbound，要按mac或
ip去设，如果你有proxy的话可以试着给packet加tag，然后
按tag设，反正我这里是按ip设的，用着没啥问题 |
|
k***e
发帖数: 7933 | 15 我刷了1.4.3v3-squashfs
可以看cctv1。但是老问题还在，有时卡。我觉得应该是源的问题，不卡的时候
bandwidth也才用到100~200KB/s，我家的internet有600+KB/s的下载bandwidth。
看youtube很流畅。 |
|
|
s******v
发帖数: 4495 | 17 what happens if you turn off other pc. if you have p2p software, they will
consume a lot of bandwidth in background.
what's the bandwidth of your plan? netflix uses less than 1m, at most 4m. |
|
W*****Y
发帖数: 433 | 18 看到网上有人提到：
“x58 mobo has a shitty Marvell 9128 SATA3 controller”
看来买SSD还得注意一下自己主板的SATA controller，x58主板上的好像都不怎么样啊
“Maximus 4 Extreme has the Marvell 9182 Chipset, while I am stuck with the
9128 Chipset on my P6X58D-e...
9128 is SATA6G controller with RAID function that only utilize the maximum
bandwidth of one single PCIEx1 lane.
9182 on the other hand is the next generation SATA 6G controller with RAID
function that utilize the maximum bandwidth of two PCIEx1 lanes. ”
原文在这里：Vertex 3 Benchmarks o... 阅读全帖 |
|
s*****e
发帖数: 404 | 19 RT-16N装的tomato vpn，家里共有3台laptop，3台desktop，还有iphone, ipad和
wireless printer。因为有一台desktop用pplive，有时占用很大带宽，所以设置QOS来
控制pplive。我的设置如下。
outbound max bandwidth：1700kbs，dslreports上实测值的85%
default class。 class c 1%-10% 17-170kbit/s
inbound max bandwidth：5000kbit/s
class c：10% 500k，
classification： High www， Low www（512k+），highest DNS，lowest DNS(2k+)
pplive ip: class C
这应该会把pplive限制在class C，最高upload 170kbit/s，下载小于500k，无论是
default class还是ip限制。但实际中，pplive的下载可以超过2000k。我有什么地方设
置有问题吗？
还有一个问题，因为work from home需要... 阅读全帖 |
|
y*****c
发帖数: 57 | 20 Bandwidth和throughput之间有区别。wifi有很多payload，你那个500Mbps是raw的
bandwidth，用户data的throughput不可能达到500Mbps的 |
|
c********l
发帖数: 8138 | 21 (NEW) ChicagoVPS - 2048MB RAM, 50GB Disk, 2TB Bandwidth, 2x IP address $40/
year - 4 Location Choices
Just in: ChicagoVPS is now offering a 2GB VPS for $40 per year, this time
with 2 IPv4 addresses and now offer an additional location option (Atlanta)
This is not a repost, just read this deal on another forum I frequent. This
is a new deal now offered in their new Atlanta location and comes with 2 IP
addresses.
I have a few VPS with these guys since last year or so and have absolutely
no complai... 阅读全帖 |
|
a*****a
发帖数: 1429 | 22 当然路由器好些的话，机器好些的话还是有不同的
RT-AC68W和rMBP，隔两堵墙：
2.4G：
./iperf -c 192.168.1.222 -i 10 -t 60
------------------------------------------------------------
Client connecting to 192.168.1.222, TCP port 5001
TCP window size: 128 KByte (default)
------------------------------------------------------------
[ 4] local 192.168.1.17 port 54891 connected with 192.168.1.222 port 5001
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 100 MBytes 84.3 Mbits/sec
[ 4] 10.0-20.0 sec 115 MBytes 96... 阅读全帖 |
|
a*****a
发帖数: 1429 | 23 Beamforming在我这里的区别
disable:
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 120 MBytes 101 Mbits/sec
[ 4] 10.0-20.0 sec 123 MBytes 103 Mbits/sec
[ 4] 20.0-30.0 sec 121 MBytes 102 Mbits/sec
[ 4] 30.0-40.0 sec 122 MBytes 102 Mbits/sec
[ 4] 40.0-50.0 sec 123 MBytes 103 Mbits/sec
[ 4] 0.0-50.0 sec 610 MBytes 102 Mbits/sec
enable:
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 135 MBytes 113 Mbits/sec
[ 4] 10.0-20.0 sec 138 MBytes ... 阅读全帖 |
|
a*****a
发帖数: 1429 | 24 这是隔着两堵墙测的。
在同一间房内
三米远：
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 189 MBytes 159 Mbits/sec
[ 4] 10.0-20.0 sec 215 MBytes 180 Mbits/sec
[ 4] 20.0-30.0 sec 212 MBytes 178 Mbits/sec
[ 4] 30.0-40.0 sec 211 MBytes 177 Mbits/sec
[ 4] 40.0-50.0 sec 210 MBytes 176 Mbits/sec
[ 4] 0.0-50.0 sec 1.01 GBytes 174 Mbits/sec
将近一米远：
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 215 MBytes 180 Mbits/sec
[ 4] 10.0-20.0 sec 222 MBytes 186... 阅读全帖 |
|
|
a*o
发帖数: 19981 | 26 c:down>iperf3 -c 192.168.1.200
Connecting to host 192.168.1.200, port 5201
[ 4] local 192.168.1.155 port 51825 connected to 192.168.1.200 port 5201
[ ID] Interval Transfer Bandwidth
[ 4] 0.00-1.00 sec 112 MBytes 941 Mbits/sec
[ 4] 1.00-2.00 sec 113 MBytes 945 Mbits/sec
[ 4] 2.00-3.00 sec 112 MBytes 943 Mbits/sec
[ 4] 3.00-4.00 sec 112 MBytes 944 Mbits/sec
[ 4] 4.00-5.00 sec 112 MBytes 943 Mbits/sec
[ 4] 5.00-6.00 sec 112 MBytes ... 阅读全帖 |
|
F***Q
发帖数: 6599 | 27
dp is only needed for high bandwidth/resolution outputs.
if your GPU can't output high bandwidth signals, there is no point carrying
a DP. of course, you expect to see more DPs on high-end cards than entry
level ones. |
|
q***e
发帖数: 90 | 28 不知道怎么搞得,从国内访问本站通过irobotsoft.com越来越不行了.好像
是irobotsoft.com也越来越差,以前每到月底,就会出现:
Bandwidth Limit Exceeded
The server is temporarily unable to service your request due to the site owner
reaching his/her bandwidth limit. Please try again later.
现在倒好,本月刚过半,15日,就不行了,后来一两天又好了.但是从20日以后就一直出现上
面的信息.看样子这样下去,以后每月就头几天能用了. |
|
g******i
发帖数: 32 | 29 假设Internet有2个计算机，已知bandwidth等信息，如果一个计算机向另一个传送一个
任意大小的文件，如何估计传送的时间？
做这个计算除了bandwidth外，还需要知道什么已知条件么？简单起见，也可以假设是
基于TCP的。
如果能给出相关文献就更好了，俺自己去看：）多谢多谢 |
|
S*A
发帖数: 7142 | 30 你确定是 WRT54S 而不是 WRT54GS? WRT54GS V4 以下都很好。
我用过 QoS, 还挺好用的。我用 openwrt. 单纯点 Web UI 也可以。
家里用主要是 uploading bandwidth limit.
limit download bandwidth 没有什么意义。
WRT54GS 还不错，我家有两个。其中一个用来开发智能自动浇水系
统就，接近可以投入使用了。可以自己上网调整比较理想的浇水
方式。
WRT54GS 如果用 tftp 方式刷的话风险还算很小的。因为用硬件
reset 按钮可以进入 tftp mode. 就算刷坏了还可以把原来的刷回来。
风险一般是那种没有硬件按钮进入 tftp mode 的，那就一定要保证有
软件的 boot wait. 不然砖头了要救回来就很麻烦，要焊串口 console
什么的。
我刷过无数回了，从来没有砖头过。 |
|
a******e
发帖数: 982 | 31 技術上還不錯， 除了還要寫 sql-like 的GQL外, 還可以接受。
但是如果是要認真的作大website， 將來費用會多少很難估計， 免費
的offering 有限制
4.1. Subject to Section 12 of the Terms, the Service is provided to you
without charge for up to 500MB of disk storage, 10GB incoming bandwidth per
day, 10GB outgoing bandwidth per day, 200 million megacycles of CPU per day
and 2,000 emails per day. These quotas will be know as the "Fee Threshold".
Google reserves the right to change this quota at its discretion.
要是將來 traffic 超過這個 Fee Threshold, 但是 Fee 又很貴怎麼辦？ |
|
m******t
发帖数: 635 | 32 If you read the post carefully, you've got the answer:
So that’s not a large number of machines, but the specs on those machines
typically aren’t available in the cloud, not at reasonable prices. Here
are some quick “scale up” server notes:
SQL servers have 384 GB of memory with 1.8TB of SSD storage
Redis servers have 96 GB of RAM
elastic search servers 196 GB of RAM
Tag engine servers have the fastest raw processors we can buy
Network cores have 10 Gb of bandwidth on each p... 阅读全帖 |
|
o**********e
发帖数: 18403 | 33 【 以下文字转载自 ITRelief 俱乐部 】
发信人: onetiemyshoe (onetiemyshoe), 信区: ITRelief
标 题: 斧头帮大牛：谈职场经验123 (转载)
发信站: BBS 未名空间站 (Fri Jun 13 13:45:00 2014, 美东)
【 以下文字转载自 SanFrancisco 讨论区 】
发信人: onetiemyshoe (onetiemyshoe), 信区: SanFrancisco
标 题: 斧头帮大牛：谈职场经验123 (转载)
发信站: BBS 未名空间站 (Thu Apr 10 17:19:34 2014, 美东)
发信人: onetiemyshoe (onetiemyshoe), 信区: ITRelief
标 题: 斧头帮大牛：谈职场经验123
发信站: BBS 未名空间站 (Thu Apr 10 14:26:18 2014, 美东)
很看好！
发信人: rayman (rayman), 信区: Seattle
标 题: 敢冒西版版规“大不惟”，就是要讨论阿三 ------职场经验谈123
发信站: BBS ... 阅读全帖 |
|
t******1
发帖数: 152 | 34 本来有些见地的贴，非要加上几句最低俗的骂人话。然后觉得被伤到的回骂。然后互骂。
就算你骂得对方哑了火，你会觉得高兴吗？你不觉得你的整个人在众人面前的形象变得
像个低俗的泼妇、与雅沾不上边了吗？“杭州师范的”可以说自己是马云，但他是不是
马云是别人说了算的。你骂得难听只能掉你自己的价。即便他真是马云也会有人说他算
不上马云，或者所有的马云都如何如何。
有十个华人的地方就有十派，这是非常令人遗憾的事。最终伤到的是包括你自己的所有
华人。我们的竞争对手是烙印也好，老白也罢，我们自己已经输了。
到了美国，我们起码应该学学美国本土的好的地方。互相多说声谢谢。觉得伤到了别人
说声对不起。发表意见前加一个“在下的拙见是（IMHO）”。谈不拢就“agree to
disagree”，说声再见。
你可以试试。心情好了可以延年益寿哦！作为新年的“resolution”怎么样？
在下对"latency"、"throughput"也有些体会。例如Pentium 4。它是靠提升运行频率来
增加"throughput"的。但它的"latency"很糟糕。举个极端的例子：假如它平均每秒可
处理三十亿个指令，应该算快... 阅读全帖 |
|
h******b
发帖数: 6055 | 35 三年用不完可以卖掉。
如果只是需要虚拟机，强烈推荐digital ocean, 除非你需要amazon那些乱七八糟的服
务。
ssd提升的一塌糊涂。还有bandwidth前几个tb免费是巨大节省。 aws的bandwidth感觉
比instance还贵。 |
|
j******a
发帖数: 100 | 36 不解释，自己看
Running MPI tests with 2 processes
logfile /opt/opa/src/mpi_apps/logs/.24Sep15175112
Running in MPIRUN mode (OpenMPI).
Running osu-micro-benchmarks-3.8-July12/mpi/pt2pt/osu_bw ...
Using hosts list: /opt/opa/src/mpi_apps/mpi_hosts
Hosts in run:
sut1
sut2
+ /usr/mpi/gcc/openmpi-1.8.2a1-hfi/bin/mpirun -np 2 --bynode --allow-run-as-
root -machinefile /opt/opa/src/mpi_apps/mpi_hosts -mca plm_rsh_no_tree_spawn
1 osu-micro-benchmarks-3.8-July12/mpi/pt2pt/osu_bw
---------------------------------... 阅读全帖 |
|
d******c
发帖数: 2407 | 37 The Landscape of Parallel Computing Research - UC Berkeley EECS
1. Old CW: Power is free, but transistors are expensive.
· New CW is the “Power wall”: Power is expensive, but transistors are “
free”. That
is, we can put more transistors on a chip than we have the power to turn on.
2. Old CW: If you worry about power, the only concern is dynamic power.
· New CW: For desktops and servers, static power due to leakage can be 40%
of
total power. (See Section 4.1.)
3. Old CW: Monolithic uniprocessors ... 阅读全帖 |
|
P*****a
发帖数: 102 | 38 在默认情况下，XP Pro 会保留一块网卡的20%的带宽。
如果你用的是宽带的话可能会超过这个限制所以需要把这个限制去掉，方法如下：
1.以管理员权限登陆
2. run gpedit.msc
3. 展开 "计算机配置"
4. 展开 "管理模版"
5. 展开 "network"
6. 展开 "QoS Packet Scheduler"
7. 右边窗口双击 "limit reservable bandwidth"
8. 选中选项 "已启用"
9. 将 "Bandwidth limit %" 的值改为0
10.重启。
Home版不支持该操作，无法去除该限制！ |
|
