Zen

Translation

The receiver may adjust its advertised receive window based on its current memory consumption, but also to limit the bandwidth used by the sender. In practice, the receive buffer can also shrink as the application may not able to process the received data quickly enough. In this case, the receive buffer may be completely full and the advertised receive window may shrink to `0`. When the sender receives an acknowledgment with a receive window set to `0`, it is blocked until it receives an acknowledgment with a positive receive window. Unfortunately, as shown in the figure below, the loss of this acknowledgment could cause a deadlock as the sender waits for an acknowledgment while the receiver is waiting for a data segment.
0/7310 · 731

Sign in to save translations.

English Czech Actions
A second scenario is when the initiating entity sends a `CR` segment that does not reach the remote entity and receives a duplicate `CA` segment from a previous connection attempt. This duplicate `CA` segment cannot contain a valid acknowledgment for the `CR` segment as the sequence number of the `CR` segment was extracted from the transport clock of the initiating entity. The `CA` segment is thus rejected and the `CR` segment is retransmitted upon expiration of the retransmission timer.
 
Three-way handshake : recovery from a duplicate `CA`
 
Three-way handshake : recovery from duplicates `CR` and `CA`
 
Data transfer
 
Now that the transport connection has been established, it can be used to transfer data. To ensure a reliable delivery of the data, the transport protocol will include sliding windows, retransmission timers and `go-back-n` or `selective repeat`. However, we cannot simply reuse the techniques from the datalink because a transport protocol needs to deal with more types of errors than a reliable protocol in datalink layer. The first difference between the two layers is the transport layer must face with more variable delays. In the datalink layer, when two hosts are connected by a link, the transmission delay or the round-trip-time over the link is almost fixed. In a network that can span the globe, the delays and the round-trip-times can vary significantly on a per packet basis. This variability can be caused by two factors. First, packets sent through a network do not necessarily follow the same path to reach their destination. Second, some packets may be queued in the buffers of routers when the load is high and these queuing delays can lead to increased end-to-end delays. A second difference between the datalink layer and the transport layer is that a network does not always deliver packets in sequence. This implies that packets may be reordered by the network. Furthermore, the network may sometimes duplicate packets. The last issue that needs to be dealt with in the transport layer is the transmission of large SDUs. In the datalink layer, reliable protocols transmit small frames. Applications could generate SDUs that are much larger than the maximum size of a packet in the network layer. The transport layer needs to include mechanisms to fragment and reassemble these large SDUs.
 
To deal with all these characteristics of the network layer, we need to adapt the techniques that we have introduced in the datalink layer.
 
Using sequence numbers to count bytes has also one advantage when the transport layer needs to fragment SDUs in several segments. The figure below shows the fragmentation of a large SDU in two segments. Upon reception of the segments, the receiver will use the sequence numbers to correctly reorder the data.
 
`Go-back-n` and `selective repeat` can be used in the transport layer as in the datalink layer. Since the network layer does not guarantee an in-order delivery of the packets, a transport entity should always store the segments that it receives out-of-sequence. For this reason, most transport protocols will opt for some form of selective repeat mechanism.
 
In the datalink layer, the sliding window has usually a fixed size which depends on the amount of buffers allocated to the datalink layer entity. Such a datalink layer entity usually serves one or a few network layer entities. In the transport layer, the situation is different. A single transport layer entity serves a large and varying number of application processes. Each transport layer entity manages a pool of buffers that needs to be shared between all these processes. Transport entity are usually implemented inside the operating system kernel and shares memory with other parts of the system. Furthermore, a transport layer entity must support several (possibly hundreds or thousands) of transport connections at the same time. This implies that the memory which can be used to support the sending or the receiving buffer of a transport connection may change during the lifetime of the connection [#fautotune]_ . Thus, a transport protocol must allow the sender and the receiver to adjust their window sizes.
 
To deal with this issue, transport protocols allow the receiver to advertise the current size of its receiving window in all the acknowledgments that it sends. The receiving window advertised by the receiver bounds the size of the sending buffer used by the sender. In practice, the sender maintains two state variables : `swin`, the size of its sending window (that may be adjusted by the system) and `rwin`, the size of the receiving window advertised by the receiver. At any time, the number of unacknowledged segments cannot be larger than :math:`\min(swin,rwin)` [#facklost]_ . The utilization of dynamic windows is illustrated in the figure below.
 
Dynamic receiving window
 
The receiver may adjust its advertised receive window based on its current memory consumption, but also to limit the bandwidth used by the sender. In practice, the receive buffer can also shrink as the application may not able to process the received data quickly enough. In this case, the receive buffer may be completely full and the advertised receive window may shrink to `0`. When the sender receives an acknowledgment with a receive window set to `0`, it is blocked until it receives an acknowledgment with a positive receive window. Unfortunately, as shown in the figure below, the loss of this acknowledgment could cause a deadlock as the sender waits for an acknowledgment while the receiver is waiting for a data segment.
 
Risk of deadlock with dynamic windows
 
To solve this problem, transport protocols rely on a special timer : the `persistence timer`. This timer is started by the sender whenever it receives an acknowledgment advertising a receive window set to `0`. When the timer expires, the sender retransmits an old segment in order to force the receiver to send a new acknowledgment, and hence send the current receive window size.
 
To conclude our description of the basic mechanisms found in transport protocols, we still need to discuss the impact of segments arriving in the wrong order. If two consecutive segments are reordered, the receiver relies on their sequence numbers to reorder them in its receive buffer. Unfortunately, as transport protocols reuse the same sequence number for different segments, if a segment is delayed for a prolonged period of time, it might still be accepted by the receiver. This is illustrated in the figure below where segment `D(1,b)` is delayed.
 
Ambiguities caused by excessive delays
 
To deal with this problem, transport protocols combine two solutions. First, they use 32 bits or more to encode the sequence number in the segment header. This increases the overhead, but also increases the delay between the transmission of two different segments having the same sequence number. Second, transport protocols require the network layer to enforce a `Maximum Segment Lifetime (MSL)`. The network layer must ensure that no packet remains in the network for more than MSL seconds. In the Internet the MSL is assumed [#fmsl]_ to be 2 minutes :rfc:`793`. Note that this limits the maximum bandwidth of a transport protocol. If it uses `n` bits to encode its sequence numbers, then it cannot send more than :math:`2^n` segments every MSL seconds.
 
Connection release
 
When we discussed the connection-oriented service, we mentioned that there are two types of connection releases : `abrupt release` and `graceful release`.
 
Abrupt connection release
 
As the entity that sends the `DR` segment cannot know whether the other entity has already sent all its data on the connection, SDUs can be lost during such an `abrupt connection release`.
 
The second method to release a transport connection is to release independently the two directions of data transfer. Once a user of the transport service has sent all its SDUs, it performs a `DISCONNECT.req` for its direction of data transfer. The transport entity sends a control segment to request the release of the connection *after* the delivery of all previous SDUs to the remote user. This is usually done by placing in the `DR` the next sequence number and by delivering the `DISCONNECT.ind` only after all previous `DATA.ind`. The remote entity confirms the reception of the `DR` segment and the release of the corresponding direction of data transfer by returning an acknowledgment. This is illustrated in the figure below.
 
Graceful connection release
 
Footnotes
 
For example, the :manpage:`htonl(3)` (resp. :manpage:`ntohl(3)`) function the standard C library converts a 32-bits unsigned integer from the byte order used by the CPU to the network byte order (resp. from the network byte order to the CPU byte order). Similar functions exist in other programming languages.
 
In the application layer, most servers are implemented as processes. The network and transport layer on the other hand are usually implemented inside the operating system and the amount of memory that they can use is limited by the amount of memory allocated to the entire kernel.
 

Loading…

User avatar None

Source string added
The receiver may adjust its advertised receive window based on its current memory consumption, but also to limit the bandwidth used by the sender. In practice, the receive buffer can also shrink as the application may not able to process the received data quickly enough. In this case, the receive buffer may be completely full and the advertised receive window may shrink to `0`. When the sender receives an acknowledgment with a receive window set to `0`, it is blocked until it receives an acknowledgment with a positive receive window. Unfortunately, as shown in the figure below, the loss of this acknowledgment could cause a deadlock as the sender waits for an acknowledgment while the receiver is waiting for a data segment.
6 years ago
Browse all string changes

Glossary

English Czech
No related strings found in the glossary.

18767 String information

Source string location
../../principles/transport.rst:769
String age
6 years ago
Last updated
2 months ago
Source string age
6 years ago
Translation file
locale/cs/LC_MESSAGES/principles/transport.po, string 144