z***e
发帖数: 5600 | 1 There is a freeware called backoff that can turns it off. Google it.
I am also wondering if there is a fast way to back up iPad. The movies sync
Very fast once backup is turned off so it is not a data transfer speed issue |
|
a***y
发帖数: 19743 | 2 ☆─────────────────────────────────────☆
wodemj (wodemj) 于 (Mon Jul 26 00:42:47 2010, 美东) 提到:
这个同步功能太烦人了，刚才一把itouch插到计算机上，就开始同步，结果把我装的
application程
序删了一半。其中还有的是我付费买的。
请问是不是就找不回来了。重新下载还要再花一次钱？
☆─────────────────────────────────────☆
zygote (yokel) 于 (Mon Jul 26 09:44:02 2010, 美东) 提到:
麻烦你有时间好好学学怎么用iTunes再来抱怨。
☆─────────────────────────────────────☆
zooie (zooie) 于 (Mon Jul 26 10:05:00 2010, 美东) 提到:
There is a freeware called backoff that can turns it off. Google it.
I am also wonderi... 阅读全帖 |
|
w********2
发帖数: 16371 | 3 现在基本上就是几个技术公司try 极限
被抓到就道歉backoff 一点。
连google 地图车偷数据都是说一句不是故意的也就没事情了
social
long |
|
z*****n
发帖数: 7639 | 4 In that case it is totally fair for 2 stations...
802 MAC works in such a way:
if a transmission is sensed when counting down the
backoff slots, halt the counting until the channel
is clear again.
if collision happens, double the contention window.
if a successful transmission confirmed, contention
window resets to minimum.
All the nodes have equal chance to access the channel,
no matter how many packets are pending in upper layer.
UDP |
|
m*****n
发帖数: 5611 | 5 这个一般是exponential backoff，没动静的话就一个比一个时间长，隔多久取决于死
了多久 |
|
Z****g
发帖数: 13731 | 6 支持，的确有点对不住楼主发工作贴的好意
叫人BACKOFF的可以另开一贴，奉陪啊 |
|
z*****n
发帖数: 7639 | 7 yes, it increases throughput quite a lot, because
a big portion of time in 802.11 MAC protocol is
wasted on catching the channel (random-backoff). |
|
|
|
z*****n
发帖数: 7639 | 10
IP layer has function of re-assembly function.
It is for that an IP packet may be fragemented when
being sent out or at a point in the network where
MTU is lower. Yes, you are right, the IP layer
reassembles an TCP packet for the upper layer.
TCP is designed to ACK every incoming packet for sure.
However, accumulative ACK is most seen (I guess) is
caused by that TCP daemon is called foreward only in
its own time slice in a multi-task OS. During its
backoff time several TCP packets may have alre |
|
s********k
发帖数: 6180 | 11 无论是contention导致的loss还是link layer的BER导致的loss。有没有办法在MAC区分
这两种不同的loss |
|
z*****n
发帖数: 7639 | 12 There is no difference between these two cases
for the sender. In both cases, the intended
receiver will not send back an ACK and the
sender's timer will go expired and resulting
in doubling the CW.
If we stick on CSMA/CA, in which the collision
is not able to be detected, MAC cannot distinguish
these two. |
|
s********k
发帖数: 6180 | 13 不知道现在的操作系统，比如xp或者linux之类的，在WLAN的情况下，TCP协议是否还是
传统的reno或者new reno，或者做了其他的优化。昨天用wireshark看了一下，好像XP
系统TCP的window并不是按照slow start，congestion avoidance来做的。window size
一直在某个值附近震荡，不知道哪位知道XP系统是否对这些做了优化 |
|
c****n
发帖数: 21367 | 14 那是因为你的机器上不止一个tcp flow吧
XP
size |
|
s********k
发帖数: 6180 | 15 昨天测试的时候就打开的一个网页。再说即使多个TCP flow，应该也有slow start啊。
不会一下congestion window size就变得很大 |
|
z*****n
发帖数: 7639 | 16 应该还是reno吧，没听说现在哪个操作系统对wireless
用了另外一套scheme。
至于你说的wireshark的capture result，那个是指的
receive window，not congestion window.
XP
size |
|
s********k
发帖数: 6180 | 17 了解。那有什么工具可以看congestion window的size |
|
|
c****n
发帖数: 21367 | 19 windows下微软有一套对学生免费的2.5层工具，也要修改代码... |
|
|
c*****s
发帖数: 184 | 21 Some hot RFf topics recently:
GaN at >Ka band, TX/RX at E/D band
GaN electrothermal model, semiconductor device model at D band
High PAE PA at xx dB backoff |
|
c*****s
发帖数: 184 | 22 Some hot RFf topics recently:
GaN at >Ka band, TX/RX at E/D band
GaN electrothermal model, semiconductor device model at D band
High PAE PA at xx dB backoff |
|
m*********k
发帖数: 10521 | 23 "[NCAA] Teamster Feb 1 ● NSD 包子"
成功扣除 4114 伪币的用户: Teamster
成功奖励 20 伪币的用户: francexplosn, JimTressel, SinoGator, leehwy, domer71
, NYPD, Bar, FIBA, edooo, scarlett, Jets, brmj, images, zkjeffrey, remrofia,
meflying, Zaini, lisk, xinlei, airjojo, fanfaner, smallforest, moneybull,
Sydra, coolbeans, o2, anxin, ccwf, myear2012, baise, cnbuff410, niuniu522,
Temudgin, wavebye, tzcyji6586, liarliar, clover4, UHL, uuuuhlhl, forgether,
jadszlf6, islands, godorme, wh, cuoai, pplulu, Grad, spox, missut... 阅读全帖 |
|
m*********k
发帖数: 10521 | 24 "[Geisha ]Geisha Feb 2 ● [bssd]谢谢各位支持, Geisha顺利当选十大优秀
版"
成功扣除 3300 伪币的用户: Geisha
成功奖励 10 伪币的用户: ZQL, ingoodmood, mypeak, XOXO2012, oriolechao01,
phox22919, jiawen, leooooo, dzhlh, happigee, PInce, hunzaimeiguo, omo, kkbb,
Go8, Orange23, Haoyun2011, tndb, haaaaaappier, MissA, fisherduyu, navaratna
, lxinzheng, hskevin, rainyj, hic, yiban, aplets, ruskiller, Buxinjiao,
Quill, comcom, ludeackpower, qw, yasaka, donic, stiga, goldmetal, DHA,
aablackk, ivylee16, juic, tibhar, Joola, Acousticia... 阅读全帖 |
|
m*********k
发帖数: 10521 | 25 "[Basketball]euclid2003 Feb 4. ● 散尽家财给Jeremy Lin今天的表现的发包子"
成功扣除 2387 伪币的用户: pandamalone
成功奖励 10 伪币的用户: cnnssb, malaichua, ruro, TheAnswer, BonnieBrae,
loyh5, baybear, Casillas, Jingqg1979, sdecui, HL, FIBA, sioc, MrYi, juto,
orz, roex, kaylaxu, Drake, fanfaner, killmesoftly, kanbbs, tamuer, IIl,
Acartia, question1, party, goldigger, ohbaby, hurtkua, ImHot, andysun,
Mutombo, JN, ZBo, masonrule, soulhunter, eggpain, Xnewbuilt, Hilfiger, laoer
, MoonWater, Master, IC, goodtiming, equator, wohenh... 阅读全帖 |
|
