EfficiencyEfficiencyis a term borro

Efficiency
Efficiencyis a term borrowed from engineering and scientific fields. It is a measurement
of how effective an operation is in comparison to the cost in effort, energy, time, or
money. Efficiency specifies how much overhead is required to produce a required outcome. For example, you could measure the efficiency of a method for boiling water.
Does most of the energy go to actually boiling the water or does a lot of the energy get
wasted heating the electrical wiring, the pot the water is in, and the air around it? How
much overhead is required to produce the desired outcome?
Efficiency also provides a useful way to talk about network performance. For example,
shared Ethernet is inefficient when the collision rate is high. (The amount of effort to successfully send a frame becomes considerable because so many frames experience collisions.) Network efficiency specifies how much overhead is required to send traffic,
whether that overhead is caused by collisions, token passing, error reporting, rerouting,
acknowledgments, large frame headers, a bad network design, and so on.
Large frame headers are one cause for inefficiency. We worry a lot less about frame
headers than we used to when bandwidth was scarcer. Nonetheless, for networks where
40 Top-Down Network Design
Small Frames (Less Efficient)
Large Frames (More Efficient)
Figure 2-2 Bandwidth Utilization Efficiency
for Small Versus Large Frames
bandwidth is still (or may become) scarce, a good network performance goal is for
applications that send bulk data to minimize the amount of bandwidth used by headers
by using the largest possible frame the MAC layer allows. Using a large frame maximizes
the amount of useful application data compared to header data and improves application
layer throughput.
Figure 2-2 shows a bandwidth pipe used by small frames and the same pipe used by large
frames. The header of each frame is shaded. Note that there is an interframe gap between
each frame in addition to the headers. From the graphic, you can see that large frames
use bandwidth more efficiently than small frames.
The maximum frame size is a tradeoff with the BER discussed in the previous section.
Bigger frames have more bits and hence are more likely to be hit by an error. If there were
no errors, an infinitely big frame would be the most efficient (although not the most fair
to other senders). If a frame is hit by an error, it must be retransmitted, which wastes time
and effort and reduces efficiency. The bigger the frame, the more bandwidth is wasted
retransmitting. So, because networks experience errors, frame sizes are limited to maximize efficiency and fairness. The maximum frame size for Ethernet, for example, is 1522
bytes, including the header, CRC, and an 802.1Q VLAN tag.
As is the case with many network design goals, there are tradeoffs associated with a goal
of improving efficiency by using large frame sizes. On slow WAN links, the time to output a large frame is significant. The time to output a frame is called serialization delay.
Serialization delay becomes an issue when applications that send large frames, such as file
transfer, share a WAN link with applications that are delay-sensitive, such as voice and
video. One solution is to use ATM, which divides frames into cells. Other solutions
include the use of link-layer fragmentation and interleaving options, such as Frame Relay
FRF.12, Multilink Frame Relay (FRF.16), and Multilink PPP.

Kết quả (Việt) 1:

Hiệu quảEfficiencyis cụm từ vay mượn từ các lĩnh vực khoa học và kỹ thuật. Nó là một thước đovề cách có hiệu quả một hoạt động so với chi phí trong nỗ lực, năng lượng, thời gian, hoặctiền. Hiệu quả xác định chi phí bao nhiêu là cần thiết để sản xuất một kết quả cần thiết. Ví dụ, bạn có thể đo lường hiệu quả của phương pháp đun sôi nước.Đa số năng lượng đi để thực sự sôi nước hoặc không nhận được rất nhiều năng lượnglãng phí sưởi ấm dây điện, nồi nước là trong và không khí xung quanh nó? Làm thế nàonhiều chi phí là cần thiết để sản xuất các kết quả mong muốn?Hiệu quả cũng cung cấp một cách hữu ích để nói về hiệu suất mạng. Ví dụ:Ethernet được chia sẻ là không hiệu quả khi va chạm mức cao. (Số lượng các nỗ lực để thành công gửi một khung trở nên đáng kể bởi vì rất nhiều khung kinh nghiệm va chạm.) Xác định mạng lưới hiệu quả chi phí bao nhiêu là cần thiết để gửi lưu lượng truy cập,cho dù chi phí đó gây ra bởi va chạm, token passing, lỗi báo cáo, chuyển hướng,acknowledgments, tiêu đề lớn khung, một thiết kế xấu mạng, và như vậy.Lớn khung tiêu đề là một trong những nguyên nhân không hiệu quả. Chúng tôi lo lắng nhiều hơn về khungtiêu đề hơn, chúng tôi sử dụng để khi băng thông đã scarcer. Tuy nhiên, cho các mạng nơiThiết kế 40 top-Down mạngNhỏ khung (kém hiệu quả)Lớn khung (hiệu quả hơn)Hình 2-2 băng thông sử dụng hiệu quảcho nhỏ so với khung hình lớnbandwidth is still (or may become) scarce, a good network performance goal is forapplications that send bulk data to minimize the amount of bandwidth used by headersby using the largest possible frame the MAC layer allows. Using a large frame maximizesthe amount of useful application data compared to header data and improves applicationlayer throughput.Figure 2-2 shows a bandwidth pipe used by small frames and the same pipe used by largeframes. The header of each frame is shaded. Note that there is an interframe gap betweeneach frame in addition to the headers. From the graphic, you can see that large framesuse bandwidth more efficiently than small frames.The maximum frame size is a tradeoff with the BER discussed in the previous section.Bigger frames have more bits and hence are more likely to be hit by an error. If there wereno errors, an infinitely big frame would be the most efficient (although not the most fairto other senders). If a frame is hit by an error, it must be retransmitted, which wastes timeand effort and reduces efficiency. The bigger the frame, the more bandwidth is wastedretransmitting. So, because networks experience errors, frame sizes are limited to maximize efficiency and fairness. The maximum frame size for Ethernet, for example, is 1522bytes, including the header, CRC, and an 802.1Q VLAN tag.As is the case with many network design goals, there are tradeoffs associated with a goalof improving efficiency by using large frame sizes. On slow WAN links, the time to output a large frame is significant. The time to output a frame is called serialization delay.Serialization delay becomes an issue when applications that send large frames, such as filetransfer, share a WAN link with applications that are delay-sensitive, such as voice andvideo. One solution is to use ATM, which divides frames into cells. Other solutionsinclude the use of link-layer fragmentation and interleaving options, such as Frame RelayFRF.12, Multilink Frame Relay (FRF.16), and Multilink PPP.