不要迷恋哥，哥只是一个传说
[置顶] 关于本blog的声明
2008/11/20 12:36 | 
很久没有写过日志，提着笔一直很生疏，也一直不想写什么，有时候觉得过去了就过去了，不想写什么，也怕以后看到万一会伤感。
24号下午1点55的飞机，虽然是个什么鸟祥鹏航空，不过第1次坐飞机，还是相当兴奋的。起飞和降落的时候才有一点点感觉，突然的失重和超重，很喜欢那样的感觉，耳膜也再一次提出了抗议，上次在武当山做缆车，就是因为短时间内海拔降低气压变化而导致耳膜疼痛，这次也同样的光临了我。就那样疼了2小时15分终于到了大连周水子机场。
跟着Neusoft的车来到了东软，开始了一个月城堡式的监狱生活。整个Neusoft的软件园区城堡式的建筑，拍了很多Photo，由于这鸟地方，USB口给ata; name="tags"
24号下午1点55的飞机，虽然是个什么鸟祥鹏航空，不过第1次坐飞机，还是相当兴奋的。起飞和降落的时候才有一点点感觉，突然的失重和超重，很喜欢那样的感觉，耳膜也再一次提出了抗议，上次在武当山做缆车，就是因为短时间内海拔降低气压变化而导致耳膜疼痛，这次也同样的光临了我。就那样疼了2小时15分终于到了大连周水子机场。
跟着Neusoft的车来到了东软，开始了一个月城堡式的监狱生活。整个Neusoft的软件园区城堡式的建筑，拍了很多Photo，由于这鸟地方，USB口给ata; name="tags"
很久没有写过日志，提着笔一直很生疏，也一直不想写什么，有时候觉得过去了就过去了，不想写什么，也怕以后看到万一会伤感。
24号下午1点55的飞机，虽然是个什么鸟祥鹏航空，不过第1次坐飞机，还是相当兴奋的。起飞和降落的时候才有一点点感觉，突然的失重和超重，很喜欢那样的感觉，耳膜也再一次提出了抗议，上次在武当山做缆车，就是因为短时间内海拔降低气压变化而导致耳膜疼痛，这次也同样的光临了我。就那样疼了2小时15分终于到了大连周水子机场。
跟着Neusoft的车来到了东软，开始了一个月城堡式的监狱生活。整个Neusoft的软件园区城堡式的建筑，拍了很多Photo，由于这鸟tion: form-data; name="summaryway"
0
24号下午1点55的飞机，虽然是个什么鸟祥鹏航空，不过第1次坐飞机，还是相当兴奋的。起飞和降落的时候才有一点点感觉，突然的失重和超重，很喜欢那样的感觉，耳膜也再一次提出了抗议，上次在武当山做缆车，就是因为短时间内海拔降低气压变化而导致耳膜疼痛，这次也同样的光临了我。就那样疼了2小时15分终于到了大连周水子机场。
跟着Neusoft的车来到了东软，开始了一个月城堡式的监狱生活。整个Neusoft的软件园区城堡式的建筑，拍了很多Photo，由于这鸟tion: form-data; name="summaryway"
0
2009/07/02 15:26 |
排线顺序有两类，A类和B类, B类是橙白、橙、绿白、蓝、蓝白、绿、棕白、棕。A类是绿白、绿、橙白、蓝、蓝白、橙、棕白、棕。
而你的网线制作的时候要选择A类还是B类，这个要看你的网络环境，一般电脑直接接到电脑之类的对等网，网线一头是A类，另一头是B类，如果是电脑接到交换机或者hub则两头都是B类或者是A类，一般习惯上多数都是做成B类。
另外需要注意的是，有些朋友做网线不喜欢按照标准来做，而是喜欢按照自己的习惯来排列顺序，这样做虽然网线也会通，但是可能会影响网络数据传输的速率，网线之所以要用双绞线是因为它的形状可以有效的减少干扰。如果顺序错了可能会使这种抗干扰能力下降。
而你的网线制作的时候要选择A类还是B类，这个要看你的网络环境，一般电脑直接接到电脑之类的对等网，网线一头是A类，另一头是B类，如果是电脑接到交换机或者hub则两头都是B类或者是A类，一般习惯上多数都是做成B类。
另外需要注意的是，有些朋友做网线不喜欢按照标准来做，而是喜欢按照自己的习惯来排列顺序，这样做虽然网线也会通，但是可能会影响网络数据传输的速率，网线之所以要用双绞线是因为它的形状可以有效的减少干扰。如果顺序错了可能会使这种抗干扰能力下降。
Open-loop congestion control works well for certain types of applications, as we will see when we discuss bandwidth reservations later in the course. However, for many applications, the use of open-loop admission control would result in inefficient use of the network. Let us look at an example of a user who is surfing the Web. The user will typically not know in advance what pages he or she will visit or how much bandwidth is needed. When forced to request a certain amount before starting to use the network, the user is likely to pick a bandwidth that is high enough so that most Web pages can be downloaded in a responsive manner. The problem is that while the user is reading the page, this bandwidth goes unused and is wasted. For applications that generate heavy, sporadic traffic, such as Web surfing, most of the allocated bandwidth is likely to go unused.
An alternative to open-loop control is to use closed-loop control. With closed-loop control, our Web surfer can use whatever bandwidth is available when downloading pages. When the network observes the onset of congestion on a specific network link, it notifies the users who are sending traffic over that bottleneck link and asks them to reduce their transmission rate. The goal is to reduce the transmission rates sufficiently so that the sum of the bandwidths used by all the users of the link is less than (but close to) the capacity of the link. Closed-loop congestion control allows users to adapt their bandwidth use to the available capacity in a more dynamic and opportunistic way. When our Web surfer has finished downloading a page and is reading the page, the network bandwidth he or she was using is immediately available for other network users. The essential behavior that makes closed-loop congestion control work is that traffic sources reduce their transmission rate when they receive a congestion feedback signal from the network, thus reducing the amount of traffic inserted in the network to below the capacity of the network.
Closed-loop congestion control involves three separate mechanisms that have to work together — detection of the onset of congestion on the router or switch, feedback to the sources using the congested link, and adaptation of the transmission rate by the sources. We briefly discuss some possible solutions for each mechanism below.
Detection
Routers or switches can detect the onset of congestion by monitoring their packet queues. If a queue is filling up — that is, packets are added to the queue faster than they can be forwarded, it is a sign that congestion is building and that large numbers of packets may have to be dropped if no action is taken. Note that there is delay associated with closed-loop congestion control. Specifically, from the time that a router decides that congestion is building until the time that the router can observe the effects of the reduced source transmission rates, at least one packet roundtrip time between the router and the source will have elapsed. For this reason, routers may want to start sending congestion feedback to sources before the packet queue fills up completely.
Feedback
Sending feedback to the sources is a difficult problem — in part, because there may be many sources, and in part, because the router, or switch, does not always know the sources. We can distinguish between implicit and explicit feedback. The Internet today relies exclusively on implicit feedback. With implicit feedback, routers start dropping packets when congestion is starting. After a while, the transport protocol on the sending host will notice that one of the packets was not received by the destination host. It will implicitly assume that the packet loss was a result of congestion, and it will reduce its transmission rate. While providing feedback by dropping packets may seem crude, it has the advantage that it is simple and universally applicable. Any switch or router can drop packets, and in fact, a router or switch does not actively participate in congestion control, the default behavior is to drop packets (because of buffer overflow) when congestion has started. An alternative to implicit congestion control is to have routers send an explicit congestion signal to the sources that use the congested link. The simplest way of implementing explicit feedback is to have the router send a special message to the sources. While simple, this becomes expensive when there are many sources.
An alternative is to have the router or switch mark packets that travel over the congested link, for example by setting a special bit in the header. When the destination host observes marked packets, it can notify the source of the packet. Explicit congestion notification has been used in a number of data link technologies, for example in ATM. Recently, the IETF also included an Explicit Congestion Notification (ECN) field in the IP header, so we could see the use of explicit congestion notification (besides implicit congestion notification) in the Internet in the near future.
Adaptation
When a host receives a congestion feedback signal from the network, it should reduce its transmission rate so congestion conditions can be relieved. But by how much should it reduce its rate? Some theoretical work has shown that in order to avoid collapse of a network due to congestion, the source host should reduce its transmission rate in a multiplicative way when it receives a congestion notification—that is, it should multiply its rate with a number smaller than one. A typical value is 0.5. On the other hand, when it increases its rate (see below), it should do so incrementally—that is, it should add some constant amount to its rate. This means that hosts should reduce their transmission rate much more aggressively than they increase their rate. Intuitively this makes sense. When the network is in trouble (i.e., congestion is building), you would like it to get out of trouble as quickly as possible. On the other hand, when the network is fine and a source is adding load, it should do so carefully so as not to create congestion.
So far, we have focused on the question of how we can control congestion by having the traffic sources reduce their transmission rates. Of course, if hosts can only reduce their transmission rate and can never increase it, network performance is likely to be very poor. Imagine a car with only brakes but no gas pedal! Another way of posing the question is as follows: if there is unused bandwidth in the network, how can hosts learn about it so they can increase their transmission rate and benefit from the unused network bandwidth? In networks that use explicit congestion notification, it is also possible for the network to tell the hosts explicitly that more bandwidth is available. While some networks (for example, certain ATM networks) use this strategy, it is rare, in part because this solution adds complexity to the network. In most networks, it is up to the traffic sources to discover that additional network bandwidth is available. This is done using probing. Periodically, traffic sources increase their traffic rate by a small amount. If they receive a congestion notification, there is no excess bandwidth available, and they reduce their rate. If they do not receive a congestion notification, there is excess bandwidth available, and the host continues to send at the higher rate. Of course, to maintain network stability, rate increases are incremental, and rate reductions are multiplicative. As a result, we end up with a somewhat unexpected result. In a network that has a constant traffic load (that is, a fixed number of traffic sources are continuously active), the transmission rates of a source will change all the time as sources probe for more bandwidth and reduce their rate when they eventually get a congestion notification.
An alternative to open-loop control is to use closed-loop control. With closed-loop control, our Web surfer can use whatever bandwidth is available when downloading pages. When the network observes the onset of congestion on a specific network link, it notifies the users who are sending traffic over that bottleneck link and asks them to reduce their transmission rate. The goal is to reduce the transmission rates sufficiently so that the sum of the bandwidths used by all the users of the link is less than (but close to) the capacity of the link. Closed-loop congestion control allows users to adapt their bandwidth use to the available capacity in a more dynamic and opportunistic way. When our Web surfer has finished downloading a page and is reading the page, the network bandwidth he or she was using is immediately available for other network users. The essential behavior that makes closed-loop congestion control work is that traffic sources reduce their transmission rate when they receive a congestion feedback signal from the network, thus reducing the amount of traffic inserted in the network to below the capacity of the network.
Closed-loop congestion control involves three separate mechanisms that have to work together — detection of the onset of congestion on the router or switch, feedback to the sources using the congested link, and adaptation of the transmission rate by the sources. We briefly discuss some possible solutions for each mechanism below.
Detection
Routers or switches can detect the onset of congestion by monitoring their packet queues. If a queue is filling up — that is, packets are added to the queue faster than they can be forwarded, it is a sign that congestion is building and that large numbers of packets may have to be dropped if no action is taken. Note that there is delay associated with closed-loop congestion control. Specifically, from the time that a router decides that congestion is building until the time that the router can observe the effects of the reduced source transmission rates, at least one packet roundtrip time between the router and the source will have elapsed. For this reason, routers may want to start sending congestion feedback to sources before the packet queue fills up completely.
Feedback
Sending feedback to the sources is a difficult problem — in part, because there may be many sources, and in part, because the router, or switch, does not always know the sources. We can distinguish between implicit and explicit feedback. The Internet today relies exclusively on implicit feedback. With implicit feedback, routers start dropping packets when congestion is starting. After a while, the transport protocol on the sending host will notice that one of the packets was not received by the destination host. It will implicitly assume that the packet loss was a result of congestion, and it will reduce its transmission rate. While providing feedback by dropping packets may seem crude, it has the advantage that it is simple and universally applicable. Any switch or router can drop packets, and in fact, a router or switch does not actively participate in congestion control, the default behavior is to drop packets (because of buffer overflow) when congestion has started. An alternative to implicit congestion control is to have routers send an explicit congestion signal to the sources that use the congested link. The simplest way of implementing explicit feedback is to have the router send a special message to the sources. While simple, this becomes expensive when there are many sources.
An alternative is to have the router or switch mark packets that travel over the congested link, for example by setting a special bit in the header. When the destination host observes marked packets, it can notify the source of the packet. Explicit congestion notification has been used in a number of data link technologies, for example in ATM. Recently, the IETF also included an Explicit Congestion Notification (ECN) field in the IP header, so we could see the use of explicit congestion notification (besides implicit congestion notification) in the Internet in the near future.
Adaptation
When a host receives a congestion feedback signal from the network, it should reduce its transmission rate so congestion conditions can be relieved. But by how much should it reduce its rate? Some theoretical work has shown that in order to avoid collapse of a network due to congestion, the source host should reduce its transmission rate in a multiplicative way when it receives a congestion notification—that is, it should multiply its rate with a number smaller than one. A typical value is 0.5. On the other hand, when it increases its rate (see below), it should do so incrementally—that is, it should add some constant amount to its rate. This means that hosts should reduce their transmission rate much more aggressively than they increase their rate. Intuitively this makes sense. When the network is in trouble (i.e., congestion is building), you would like it to get out of trouble as quickly as possible. On the other hand, when the network is fine and a source is adding load, it should do so carefully so as not to create congestion.
So far, we have focused on the question of how we can control congestion by having the traffic sources reduce their transmission rates. Of course, if hosts can only reduce their transmission rate and can never increase it, network performance is likely to be very poor. Imagine a car with only brakes but no gas pedal! Another way of posing the question is as follows: if there is unused bandwidth in the network, how can hosts learn about it so they can increase their transmission rate and benefit from the unused network bandwidth? In networks that use explicit congestion notification, it is also possible for the network to tell the hosts explicitly that more bandwidth is available. While some networks (for example, certain ATM networks) use this strategy, it is rare, in part because this solution adds complexity to the network. In most networks, it is up to the traffic sources to discover that additional network bandwidth is available. This is done using probing. Periodically, traffic sources increase their traffic rate by a small amount. If they receive a congestion notification, there is no excess bandwidth available, and they reduce their rate. If they do not receive a congestion notification, there is excess bandwidth available, and the host continues to send at the higher rate. Of course, to maintain network stability, rate increases are incremental, and rate reductions are multiplicative. As a result, we end up with a somewhat unexpected result. In a network that has a constant traffic load (that is, a fixed number of traffic sources are continuously active), the transmission rates of a source will change all the time as sources probe for more bandwidth and reduce their rate when they eventually get a congestion notification.
也许是拜某人“饭否门”所赐，第1次觉得在网上发东西是那么的可怕，而且是根本没有公开自己的资料，信息，一直在想“饭否门”男猪脚怎么找到粥的，也一直想不通他为撒还会跟粥她妈妈说。事后把饭否停了，然后也害怕的不敢在网上写些东西，害怕被人看见，害怕引起不必要的矛盾。
昨天上3G门户，看到了一则类似艳照门的事情，某男将手机遗失后，去公安报警，希望公安通过移动公司找到该手机下落，怎么劳师动众的只是因为该手机上有某男偷拍他女友的“艳照”。
也许现在这些的隐私永远只是保留在自己的心里，就算写出来发泄完了，还是要立马删掉的，永远不能忽视了google baidu 的强大。
昨天上3G门户，看到了一则类似艳照门的事情，某男将手机遗失后，去公安报警，希望公安通过移动公司找到该手机下落，怎么劳师动众的只是因为该手机上有某男偷拍他女友的“艳照”。
也许现在这些的隐私永远只是保留在自己的心里，就算写出来发泄完了，还是要立马删掉的，永远不能忽视了google baidu 的强大。
Take Assessment: Certification Exam Practical
1. Go to bottom of question.
The SSD8v1.5 Certification Exam
Reservation System for Airline Flight Tickets
Read this entire document at least once before beginning work.
1. Go to bottom of question.
The SSD8v1.5 Certification Exam
Reservation System for Airline Flight Tickets
Read this entire document at least once before beginning work.
5月，每天基本都是复习看书，然后考试，然后继续看书，然后继续考试，仿佛永远不会停止似的。
然后感情生活又一团糟，完全不知道下一步该怎么走...
马上的6月，马上的实训，马上的考研路...
一步一步慢慢走吧
然后感情生活又一团糟，完全不知道下一步该怎么走...
马上的6月，马上的实训，马上的考研路...
一步一步慢慢走吧
