漏桶问题
我一直在尝试解决以下数值问题......感谢任何帮助使概念清晰。
“给定的源请求准入 QoS 网络，请求平均吞吐量为 2 Mbits/秒，突发容量为 2 Mbits。来源 然后以 50 Mbits/sec 的速度传输数据，持续时间为 1 毫秒。就在之后 源将吞吐量降低至 1.8 Mbits/sec。 绘制为此流保留的缓冲区中数据的大小，作为以下函数 时间与上述吞吐量并列。多少 该源会丢失数据吗？这个源的突发容量是多少 应该使用吞吐量函数显示来确保没有数据丢失 上面？”
谢谢
I have been trying to solve the following numerical problem........Any help is appreciated in making the concept clear.
"A given source request admission to a QoS network requesting avg throughput of
2 Mbits/sec and burst capacity of 2 Mbits. The source
then transmitt data at 50 Mbits/sec for a duration of 1 millisecond. Right after
that the source scales down the throughput to 1.8 Mbits/sec.
Plot the size of the data in the buffer reserved for this flow as a function of
time side by side with the throughput described above. How much
data loss will this source experience? What is the burst capacity this source
should use to ensure no data loss with throughput function show
above?"
Thank-you
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
假设
在 T = 1 mS 时，已经过去了 10^-3 秒，因此 2 *10^3 字节已从缓冲区中清除。然而，此时，客户端在 1 毫秒的持续时间内每秒吐出 (50*10^6) 字节，即总共 50*10^3 字节。
由于可用内存只有 2*10^3 字节，前 2*10^3 字节将“离线”正确读取，其余（48*10^3 字节）将丢失，或导致致命的缓冲区溢出。
如果要避免数据丢失，某处至少需要另外 48*10^3 字节的内存。与此数据突发相关，问题陈述的其余部分毫无意义，因为问题似乎是在询问支持给定突发所需的缓冲，而这是给定图表上的峰值数据速率。
我不确定您正在寻找的答案是什么，但我希望对网络机制的描述有所帮助。
Assume
At T = 1 mS, 10^-3 seconds have elapsed, so 2*10^3 bytes have been cleared from the buffer. However, in that time, the client has spat out (50*10^6) bytes/sec for the 1 mS duration, or a total of 50*10^3 bytes.
As the available memory is only 2*10^3 bytes, the first 2*10^3 bytes will read correctly "off the wire", the rest (48*10^3 bytes) will be lost, or cause a fatal buffer overflow.
Somewhere, there needs to be AT LEAST another 48*10^3 bytes of memory if data loss is to be avoided. In relation to this data burst, the rest of the problem statement is meaningless, because the question appears to be asking about the buffering required to support the given burst, and this is the peak data rate over the given graph.
I'm not sure what the answer you are seeking is, but I hope this description of the network mechanics is helpful.