Q-CSMA

1.Effect of Access Probabilities on the Delay Performance of Q-CSMA Algorithms

本文提出基于排队的CSMA算法优化，这里面涉及到两个参数，一个是transmission probability：transmission probability of each link is chosen as an appropriate function of its queuelength,另一个是access probability：are simply regarded as some random numbers since they do not play any role in establishing the network stability.本文采用access probability control CSMA马尔科夫链，得到一个公式relate the mixing time to access probabilities and use these to develop the following guideline for choosing access probabilities: each link i should choose its access probability equal to 1/(di + 1), where di is the number of links which interfere with link i.
文献的优劣：

1) every station is saturated (i.e., always has a packet waiting to be
transmitted);
2) the packet collision probability is constant
3) transmission error is a result of packets colliding and is not caused by medium errors.

Provided that every station is indeed saturated, the resulting model is remarkably accurate.
但是现实中Data traffic类似web e-mail具有突发性，streaming traffic 具有low rates andon-off manner。
因此Hence, for most real traffic the demanded transmission rate is variable with significant idle periods。那么肯定不是饱和的数据。

2.Modeling the 802.11 Distributed Coordination Function in Nonsaturated Heterogeneous Conditions
这篇文章主要考虑了不饱和的traffic assumption。进行csma/ca 建模。放松了Bianchi对饱和数据的影响。

There are approaches to nonsaturated modeling other than ours.
In [2]引入了当STA无包可穿的概率, a modification of [1] is considered where a probability of not ransmitting is introduced that represents a station having no data to send. The model is not predictive as this robability is not known as a function of load and must be estimated from simulation.

In [3], 引入了空闲状态idle states are added after packet transmission to represent bursty arrivals. The number of idle states is distributed geometrically with a parameter ; however, no relationship is given between and the load on the system. This model also includes a full backoff before each packet transmission, which does now allow for packet inter-arrival and 802.11’s post-backoff period to overlap. This model also considers multi-rate transmissions.

In [4], a Markov model where states are of fixed real-time length is introduced. As observed in the paper, the derived throughput is a monotonic function of offered load, and so the model cannot predict a pre-saturation peak in throughput.

In [5], a model focusing on multirate transmission is presented, including an infinite queue with Poisson arrivals. This model is not solved analytically and is subject to limited validation.

In [6], a non-Markov model is developed, but it is based on an unjustified assumption that the saturated setting provides good approximation to certain unsaturated quantities. It appears to produce inaccurate predictions.

None of these previous models have considered fairness issues arising from different traffic arrival rates.

The -persistent approach of the 802.11 MAC has also been studied extensively; for recent work, see [7] and the references therein.
文章要点：
1.本文主要考虑的是非饱和相同STA条件下的DCF分析，这是与Bianchi文章最大的不同。
文章依然把DCF模型建模为二维马尔科夫模型，但是增加了一个状态，一个概率–
这个状态称为（0，k）e，表示的是STA无需等待直接发送packet，文章称为postbackoff；
增加的概率体现了不饱和，即packet是以q的概率到达，这个q在文章中考虑了3种分布：
均匀分布、poisson分布和条件分布，后续实验主要考虑poisson分布。
缺点：文章中重传的原因是冲突，没有考虑噪声的影响。
3.Performance analysis under finite load and improvements for multirate 802.11

这篇文章总的来说，是对我比较有启发的文章，文中提到的关于对DCF中二进制随机回退的建模是非常清晰的。
文章考虑了两种模型：不经历回退直接从（0，w0-1）中均分分布的选择一个k值进行回退，当回退到0时，
由于packet的到来时不饱和的假设为是泊松到达，注意这里到达的时间间隔是负指数分布的，所以一个packet到达的概率是
q=1-exp（-lamda*t），
当回退到0的时候回面临一个问题，就是buffer中是否有packet（以q的概率有packet）
那么有packet，进入回退阶段（这里似乎有一点问题就是会重复一个一阶段回退过程）；没有packet进入等待阶段，在等待阶段的话等到有packet到来时，
又面临两个分类：如果信道仍然空闲，此时可以发送packet了，如果信道busy则进入回退过程。

4.A Simple Model for Nonsaturated IEEE 802.11 DCF Networks

这篇文章对之前的马氏链做了一个简化，他认为不饱和和饱和之间存在一个关系，而这个关系就是数据包的到达率，

这个大大简化了后续了工作。
5.A Queueing Model for Finite Load IEEE 802.11 Random Access MAC.
这篇文章分析了DCF的queueing delay，把每一个node建模成G/G/1队列，然后得到一个关于delay和队长的闭式表达式，文章考虑了冲突回退和packet size的分布。这篇文章对回退窗口进行了建模得到了一个平均窗口值，依据平均回退窗口值可以得到一个平均回退时间。对于回退窗口的考虑这里是我知道的第一次。

对于冲突概率

6.Performance analysis of IEEE 802.11 MAC protocols in wireless LANs
讨论了service time
3~9讨论的saturated
2提出了idea
3489分析了时延和吞吐量；4
3. P. Chatzimisios, V. Vitas and A.C. Boucouvalas, Throughput and delay analysis of IEEE 802.11 protocol,
IEEE INFOCOM’02, Vol. 1 (2002) 168–174.
4. Y.T. Lee, D.H. Han and C.G. Park, Saturation throughput and dealy analysis of DCF in the IEEE 802.11
Wireless LAN, WSEAS Trans. on Commun. 4 (2005) 776–783.
用的Bianchi的方法；3维马氏链
5. I. Tinnirello and S. Choi, Temporal fairness provisioning in multi-rate contension-based 802.11 WLANs, in
International Symposium on Autonomous Decentralized Systems, ISADS 2005, (2005) pp. 220–230.
6. H. Zhai, Y.G. Kwon and Y. Fang, Performance analysis of IEEE 802.11 MAC protocols in wireless LANs,
Wireless Commun. and Mobile Computing 4 (2004) 917–931.
涉及到M/G/1/k，涉及到服务时间
7. G. Bianchi, Performance analysis of the IEEE 802.11 distributed coordination function, IEEE J. on Selected
Areas in Commun. 18 (2000) 535–547.
2维马尔科夫链
8. P. Chatzimisios, A.C. Boucouvalas and V. Vitas, Packet dealy analysis of IEEE 802.11 MAC Protocol, Electronics
Letters 39 (2003) 1358–1359.
9. H. Chen and Y. Li, Analytical analysis of hybrid access mechanism of IEEE 802.11 DCF, IEICE Trans. on
Commun. E87-B (2004) 3682–3694.

1. G. Kramer, B. Mukherjee, X. Dixit Y. Ye and R. Hirth, Supporting differentiated class of service in ethernet
   passive optical networks, J. Optical Networking 1 (2002) 280–298.
2. C.G. Park, D.H. Han and B. Kim, Packet delay analysis of dynamic bandwidth allocation scheme in an ethernet
   10.11讨论到packet size

PON, LNCS 3420 (2005) 161–168.
12. D.H. Han, C.G. Park and K.H. Baik, Performance analysis of an AAL multiplexer with dynamic bandwidth
allocation in an IP/ATM environment, Computer & Operations Research 30 (2003) 1851–1864.
说明了采用poisson的假设不是合理额的，因为IP均有突发性和不可预知性。因此采用马尔科夫调制过程

1. A. Baiocchi and N. Blefari-Malzzi, Steady-state analysis of the MMPP/G/1/K queue, IEEE Trans. on Commun.
   41 (1993) 531–534.
2. C. Blondia, The N/G/1/K finite capacity queue, Commun. Statist. - Stochastic Models 6 (1989) 273–269.
3. C.G. Park, H.S. Jung and D.H. Han, Queueing analysis of IEEE 802.11 MAC protocol in Wireless LAN, in
   5th International Conference on Networking, ICN 2006 (Mauritius, 2006).