-Reliable Broadcast A Probabilistic Measure of Broadcast Reliability

-Reliable Broadcast:A Probabilistic Measure of Broadcast Reliability

Patrick Th.Eugster

Sun Microsystems

CH-8604Volketswil,Switzerland Rachid Guerraoui Petr Kouznetsov Distributed Programming Laboratory,EPFL CH-1015Lausanne,Switzerland

Abstract

This paper introduces a new probabilistic speci?cation of reliable broadcast communication primitives,called-Reliable Broadcast.This speci?cation captures in a precise way the reliability of practical broadcast algorithms that, on the one hand,were devised with some form of reliabil-ity in mind but,on the other hand,are not considered reli-able according to“traditional”reliability speci?cations. We illustrate the use of our speci?cation by precisely measuring and comparing the reliability of two popular broadcast algorithms,namely Bimodal Multicast and IP Multicast.In particular,we quantify how the reliability of each algorithm scales with the size of the system.

1.Introduction

The growing interest in peer-to-peer computing has un-derlined the need for reliable broadcast algorithms deploy-able at large scale.Traditionally,the reliability of broadcast algorithms has been de?ned by three properties[6]: Integrity For any message,every correct process de-livers at most once,and only if was previously broadcast by sender(m).

Validity If a correct process broadcasts a message, then eventually delivers.

Agreement If a correct process delivers a message,then every correct process eventually delivers.

To obtain these strong properties in a system with pro-cess and link failures,one employs costly,traditionally acknowledgement-based algorithms.These can be effective in a local environment,but may give unstable or unpre-dictable performance under stress,and hence tolerate lim-ited scalability.

use of probabilities enables the capture,to a certain extent, of the nondeterminism inherent to large-scale systems. We illustrate the use our measure through two well-known examples.The?rst one,Bimodal Multicast[2],is a representative of the rapidly proliferating family of gossip-based algorithms which have received much attention lately, precisely because they are“pretty reliable”.As a repre-sentative of the class of best-effort algorithms often used in practice,namely the network-level protocols,we dis-cuss IP Multicast[3]on top of which many other“reliable”network-level broadcast protocols are built.

We also demonstrate the use of-Reliability in com-paring broadcast algorithms by contrasting Bimodal Multi-cast and IP Multicast,con?rming that,in most practical en-vironments,Bimodal Multicast is“more reliable”than IP Multicast,especially as the system grows in size.This is in-sofar unsurprising as IP Multicast has not been designed to be reliable,yet illustrates the usefulness of our speci?ca-tion in.quantifying the difference between algorithms. The practical use of our-Reliability measure is further-more illustrated through the scalability analysis of Bimodal Multicast which illuminates very attractive scalability prop-erties of the algorithm.

Roadmap.Section2introduces-Reliability.Section3 discusses the-Reliability of Bimodal Multicast.Section 4similarly applies our speci?cation of-Reliability to IP Multicast.Section5illustrates the use of-Reliability in comparing broadcasting algorithms through Bimodal Mul-ticast and IP Multicast.Section6concludes with?nal re-marks,also on the applicability of our speci?cation.

2.-Reliable Broadcast:speci?cation

This section presents our approach to measuring,in a probabilistic sense,the reliability of a broadcast algorithm. (Alternatives are discussed in[4].)

2.1.System and environment

We consider an asynchronous(in the sense of[6])sys-tem of processes.Processes are connected through fair lossy channels of in?nite capacity.Let be any message,uniquely identi?ed and equipped,in partic-ular,with a parameter.Processes communi-cate by message passing de?ned by the primitives

and.Broadcast is de?ned by the primitives

and.Processes are subject to crash failures.A correct(in a given algorithm run)process is one that never crashes(in that run).To simplify presenta-tion,we do not consider Byzantine failures,and we assume that crashed processes do not recover.

The analysis of a broadcast algorithm usually depends on more properties of the underlying system than only its size and composition,as well as on parameters of the al-gorithm itself.Henceforth,we will use the term environ-ment,denoted,to refer to the set of relevant system prop-erties and algorithm parameters.Environment represents a point in an environment space,a set of all possible com-binations of parameters:.

Let and be two broadcast algorithms that have different sets of parameters in their respective environments and.To compare the algorithms we introduce a com-pound environment-a union of the two environments, .Note that the composition makes sense only if the related parameters in and do not contradict.For example,if the system models for and comprise the probabilities of an end-to-end message loss,respectively, and,then.Otherwise,the compar-ison does not seem meaningful.In Section5we will illus-trate this through the concrete examples.

2.2.-Reliable Broadcast

Let be any pair of real numbers(). We say that a broadcast protocol complies with the speci-?cation of-Reliable Broadcast(or a broadcast protocol is-Reliable)iff the following properties are simultane-ously satis?ed with probability:

Integrity For any message,every correct process de-livers at most once,and only if was previously broadcast by.

Validity If a correct process broadcasts a message then eventually delivers.

-Agreement If a correct process delivers a message, then eventually at least a fraction of correct processes deliver.

Properties Validity and Integrity here are the same as in traditional Reliable Broadcast[6].

Agreement,as de?ned in[6],is transformed here into-Agreement which is less restrictive in terms of the number of processes that need to deliver the message and also has a probabilistic?avor.

2.3.Interpretation of and

=()represents a basic“reliability measure”of a broadcast algorithm.The values of and are intrinsically coupled:can roughly be pictured as the probability with which at least a fraction of processes behave according to the properties of Reliable Broadcast[6]:

Reliability probability:is the probability that a proto-col run behaves“properly”.That is,once a message

is broadcast by a correct process,“enough”correct processes eventually deliver.

Reliability degree:de?nes the fraction of correct pro-cesses which eventually deliver.

For instance,to satisfy the properties of-Reliable Broad-cast with,once a message is broadcast,an algorithm should,with probability,de-liver to of correct processes in the system.In other terms,in a run of the system with10correct processes,one can expect95%of all messages which are broadcast to be delivered by at least9processes(not necessarily the same processes for every message).

2.4.Reliability distribution function

In a practical system,with a given required reliability degree,several broadcast algorithms can easily be com-pared along the they offer for the given.To give an in-formal measure of the general performance in terms of re-liability of a broadcast algorithm,several samples... are usually suf?cient.A precise expression of the reliabil-ity of such an algorithm requires however the consideration of the probabilities for all possible,especially when comparing two algorithms in general.Indeed,con-sider two algorithms and and a set=(0.9,0.9) and=(0.85,0.9).Algorithm seems to perform bet-ter for.However,this information is not suf?cient to promote algorithm as“more reliable”than algorithm,since for,algorithm might offer a of0.8,while in the case of algorithm, might be only0.7.

To compare two algorithms in a more general manner, we de?ne a reliability distribution function of a broadcast algorithm:

01(1) such that for any and,is-Reliable with.

As a direct consequence of the de?nition of-Agreement—a sample in which a fraction of pro-cesses deliver every message is also a sample in which at least any fraction of the processes deliver ev-ery message—()is a monotonically decreasing(with respect to)function.

Note however,that by the size of“a fraction of pro-cesses”we mean.Accordingly,()is not represented by a continuous function,but manifests steps.

https://www.wendangku.net/doc/ba10590729.html,paring broadcast algorithms Consider a reliability range,that is,a range of values for the reliability degree which is of interest in the context of a comparison.

In the sense of-Reliable Broadcast,in the environment ,an algorithm is more reliable in than an algorithm iff

and

1

(2)

Similarly,in the environment,an algorithm is said to be strictly more reliable in than an algorithm iff

(3) We exclude here,because for any broadcast algo-rithm:.

Finally,in the environment,an algorithm is more reliable than an algorithm iff,in,is more reliable than in.Analogously,in the environment, an algorithm is strictly more reliable than an algorithm iff,in,is strictly more reliable than in

.

2.6.Atomicity

The reliability distribution function can be used to de-?ne the probability that a certain number of processes de-liver the message as a result of an algorithm run.More pre-cisely,the probability that the fraction of correct processes that delivered a broadcast message(in a given environment )is larger than but smaller than() can be de?ned as:

(4) Thus,the following Atomicity predicate(see more exam-ples in[2])de?nes a failed broadcast to be one that reaches more than a fraction of correct processes,but less than a fraction of correct processes in a system().

(5) [4]discusses several alternative non-binary speci?cations of broadcast algorithms.

2.7.-Reliable Broadcast:from perfect to useless

A reliability distribution function in the sense of(1) can be found for any broadcast algorithm.We demonstrate this through the following extreme cases. Dreamcast:One can easily see that an algorithm imple-menting traditional Reliable Broadcast[6]in a given environment is-Reliable with.Since is a monotonically decreasing function,this sam-ple univocally de?nes:

.One may call such an algorithm perfectly reliable.

As we mentioned earlier in the introduction,its prac-tical implementation in a network with unreliable pro-cesses and channels is expensive and not scalable. Spellcast:A bogus algorithm which does nothing con-forms to the speci?cation of-Reliable Broad-cast such that with at least one correct process and:0().

3.Bimodal Multicast

This section focuses on the Bimodal Multicast[2]algo-rithm.While providing a lower reliability in terms of-Reliability than a perfectly reliable protocol,it is in most cases more scalable and ef?cient.We?rst recall the algo-rithm,and then discuss its-Reliability.

3.1.Protocol overview

Bimodal Multicast is composed of two subprotocols structured roughly as in the Internet MUSE protocol[7]. The?rst is an unreliable,hierarchical multicast(IP Multi-cast can be used where available)that makes best-effort at-tempt to deliver each message to its destination.The sec-ond is a two-phase anti-entropy protocol that operates in a series of asynchronous rounds.During each round,the ?rst phase detects message losses;the second phase cor-rects such losses and executes only if needed.

For the analysis below,we use a simpli?ed version of the?rst phase of the anti-entropy protocol of Bimodal Mul-ticast,which differs from the original protocol in ways that simplify the discussion without changing the analyti-cal results(also used by[2]).The algorithm proceeds as follows[2].A message which is gossiped about is at-tached the number of times it has been forwarded,. When a process receives for the?rst time,deliv-ers it,and,if the message has been forwarded less than rounds(),forwards to randomly cho-sen processes by attaching it.When a process broadcasts,it handles as if it had received with at-tached.

3.2.Model

The stochastic analysis below is based on the assumption that the execution of a broadcast algorithm can be broken up into a sequence of synchronous rounds,such that,during each round,only processes which have gossips with round number are gossiping,and every round happens strictly after all the transmission of the previous round are com-pleted[1,2].Of course,in a real execution,each process autonomously proceeds in its own asynchronous rounds.

For the following analysis,we assume that failures are stochastically independent.The probability of a message loss is,and the probability of a process crash dur-ing the protocol execution is.For simplicity,we as-sume that all incorrect processes are initially crashed.This implies that dependent link failures like a network partition are outside of our failure model.At any moment and for any message,an infected process is one that already re-ceived,an infectious process is an infected one which is gossiping in the current round,and a susceptible process is one that is not infected yet by.Following[2],we de-scribe the state of the propagation of a given message in round using the random variables,and,which de-note the number of susceptible processes and the number of infectious processes,respectively.Initially,only the broad-castin process is infected.To summarize the constraints on the state of the system:

(6)

with initial values.Note that at any round,the number of infected processes is.

3.3.Analysis

Let be the number of incorrect processes in a given run.We de?ne as the probabil-ity that a given gossip message sent by an infectious pro-cess is successfully received by a given process,that is: (a)the gossiping(infectious)process chooses as destina-tion,(b)message is not lost in transmission,and(c),pro-cess is correct.Respectively,

is the probability that a certain process did not receive a given gossip message from a particular infectious process. If processes are gossiping in a given round,susceptible process is not infected in this round with probability. The corresponding stochastic process can be expressed in the form of a homogenous Markov chain with a transition matrix de?ned by:

(7) The distribution of and can be de?ned as:

(8)

Using(6),(7)and(8),we can build a distribution of and.We are interested in the probability that,for some

,not less than a fraction of correct processes are infected up to round:

(9)

where n T is the set of system and algo-rithm parameters de?ning the current environment.

3.4.-Reliability of Bimodal Multicast

Based on this,we formally characterize the-Reliability of Bimodal Multicast[2].

Proposition1For any environment n T and any Bimodal Multicast[2]is-Reliable with

.

Proof:Validity and Integrity follow directly from the al-gorithm description and the absence of Byzantine failures: the sender of a broadcast message delivers the message im-mediately and a process that receives the broadcast message delivers it only once.Thus,Validity and Integrity are always satis?ed.

The proof of-Agreement follows from the analysis above.Since gives the probability of suc-cessfully infecting at least a fraction of correct processes, given that initially one process is infected,-Reliability with=(,)is guaranteed.

3.5.Approximation of

Here we present a way to approximate the function

.We describe the state of the system using the stochastic process-the proportion of susceptible processes in round.

Neglecting the?uctuation of around its conditional expectation,we have the following deterministic ap-proximation of the stochastic process:

(10) with the following initial conditions:

(12) Denote,where.As-sume that is constant(the number of gossip messages sent by a process per round does not depend on the sys-tem size).For large,.Thus,we have the following recursive relationship:

(14)

The question is:what is the lower-bound asymptote of the susceptible fraction of the system and how does it de-pend on?

One can easily see that equation(14)does not depend on and,that is if is a solution of(14),then,for any,is also a solution of(14).The system size only impacts the initial condition(14).Thus the lower-bound asymptote of does not depend on:(14)de?nes the time necessary to approach it.

The lower-bound asymptote can be roughly estimated for through the following consideration:

3.6.Average reliability of Bimodal Multicast

The above analysis allows to state the following result: Proposition2For any environment n in

which,the average reliability degree as a function of system size is such that:

(such that the number of partners a process gossip to each round,is constant),then the right-hand side of(17)is constant with respect to the sys-tem size.In other words,the expected reliability degree of Bimodal Multicast is stable with respect to the scale of the system.This a very valuable property for self-organizing systems,since for some?xed set of parameters of the al-gorithm,its reliability degree does not degrade as the sys-tem size increases.As we will see in the following section, IP Multicast is not scalable in this sense:its average relia-bility degree decreases exponentially as in-creases.

4.IP Multicast

In this section,we illustrate the notion of-Reliable Broadcast through a second,in the traditional sense[6]in-herently unreliable algorithm,namely IP Multicast[3].

4.1.Protocol overview

As its name reveals,IP Multicast is a so-called network-level datagram broadcast protocol directly based on IP.The transmission of such datagrams is not reliable,and basic IP Multicast does not consider message loss detection and reparation,making it inherently unreliable.In the context of IP Multicast,many different protocols have been described and deployed.

4.2.Model

We focus here on a sparse distribution of processes. We suppose a spanning tree,as for instance the ones that are encountered with the Protocol-Independent Multicast—Sparse Mode(PIM-SM)[5]protocol,which is-ary and of depth.In other terms,we consider a regular spanning tree with a single(correct)broadcasting process located at the root,receiving processes constituting the leaves of the tree,and every non-leaf node of the tree representing a router with outgoing links.The system size is thus given by,but we will consider and as pa-rameters of the environment,and,since we are interested in large systems,we use.A spanning tree ob-tained in a real use case can always be captured by a possi-bly bigger spanning tree with a number of leaves of order conforming to the above description.

Similarly to the analysis of Bimodal Multicast in the pre-vious section,is the probability that a given process fails, and the probability of a message loss in a link between two nodes in the spanning tree as.In addition,we de?ne as the probability of a router failure.We assume that all in-correct entities are initially crashed and the link failures are stochastically independent.

4.3.Analysis

Similarly to the analysis presented in the previous sec-tion,we propose a breakdown in successive rounds.These rounds however correspond to the levels in the spanning tree,that is,at round1,the router of a broadcasting pro-cess forwards a given message to the routers represent-ing its child nodes().Due to failures,only

will receive.In any round,the“infec-tious”routers of level forward to their child nodes(maximum of).The probability of a successful reception of by an entity at level is therefore given by.At round,the routers com-posing level?nally send to the processes constitut-ing the leaves of the tree.We assume that processes are correct in a given run.

The probability of having a given number of“in-fected”entities at a given level can be computed re-cursively based on the probabilities of any number of in-fected entities at level.Finally,the probability of ob-taining a given number of infected processes at the leaves of the spanning tree enables the computation of the fraction of the correct processes in which have received.For that end,we require the probability of having infected en-tities at level based on the number of infected entities at the previous level:

(18)

Thus,the probability of having infected entities at round is given recursively by:

(19)

Let be the number of incorrect processes in a given run.The probability of successful transmission of mes-sage from an infected router at level to a process at level is given by and the proba-bility of having infected processes at level based on

the number of infected entities at the previous level:

(20) Thus,the probability of having infected processes at round is given recursively by:

(21)

As a direct consequence,the probability of having infected at least a fraction of correct processes in a-ary spanning tree of depth is given by:

(22)

where is the number of correct processes in a given run and is the environment de?ned as the set of parame-ters.

4.4.-Reliability of IP Multicast

Based on(22),we are now able to formally characterize the-Reliability of IP Multicast.

Proposition3For any environment l n k and IP Multicast is-Reliable with

.

Proof:The proof of Integrity follows from the semantics of IP and the absence of Byzantine failures,and Validity is as-sured with prevalent operating systems.Thus,Validity and Integrity are always satis?ed in this model.

The proof of-Agreement follows from the analysis above.is equal to the probability of suc-cessfully infecting at least a fraction of processes.Thus -Reliability with=(,)is guaranteed.

4.5.Average reliability degree of IP Multicast The average fraction of correct processes which receive ,,is given by:

(23) Furthermore,the probability that all processes are correct and receive a given broadcast message,

,can be easily expressed in this model through:

rithms.Basically,we can consider an algorithm to be scal-able if its average reliability degree as function of the sys-tem size is constant,or slowly decreasing.

This criterion obviously re?ects just one dimension of scalability,namely that of reliability.Yet,investigating scal-ability in terms of overhead is not in the scope of this work. It is nevertheless worth noting that IP Multicast is“more scalable”in terms of message complexity and time:to ob-tain the same reliability degree it requests a smaller num-ber of messages and consumes less time.Note furthermore that traditional Reliable Broadcast[6]is scalable in this con-text:its reliability degree is,although it is not scalable in terms of message complexity and time.

5.4.Average reliability degrees

Figure2presents the average reliability degrees for Bi-modal Multicast and IP Multicast(,resp.) for a varying sytem size.As expected,does not have a noticeable impact on the reliability of Bimodal Multi-cast(see Proposition2)while,for IP Multicast,

is signi?cantly decreasing.

6.Conclusions

This paper suggests a probabilistic measure of reliabil-ity,called-Reliability.To demonstrate our measure,we considered the Bimodal Multicast algorithm of Birman et al.and a protocol variant of IP Multicast as case studies and we measured their respective reliabilities in probabilis-tic terms.The proposed speci?cations help to prove cor-rectness of other probabilistic broadcast algorithms as well as to verify upper-layer distributed computing abstractions, which are based on reliable broadcast primitives such as Bi-modal Multicast or IP Multicast.

To quantify the reliability of a broadcast algorithm in a probabilistic sense,we need the precise knowledge of sys-tem parameters and an accurate model of the behavior of the algorithm based on former ones.Such parameters are not always available,and models usually represent approx-imations.This outlines the main limitation of our notion of -Reliable Broadcast:not every system model(and algo-rithm)matches it perfectly.We understand the notions we presented here as a?rst step towards de?ning a rigorous measure for scalable and probabilistic reliable protocols. While the reliability offered by a broadcast algorithm can be quanti?ed through our approach,there is no measure of its ef?ciency so far.We are thus currently working on iden-tifying an appropriate measure for the ef?ciency,and maybe therethrough the scalability of broadcast algorithms. References

[1]N.Bailey.The Mathematical Theory of Infectious Diseases

and its Applications(second edition).Hafner Press,1975. [2]K.Birman,M.Hayden,O.Ozkasap,Z.Xiao,M.Budiu,and

Y.Minsky.Bimodal Multicast.ACM Transactions on Com-puter Systems,17(2):41–88,May1999.

[3]S.Deering and D.Cheriton.Multicast Routing in Datagram

Internetworks and Extended LANs.ACM Transactions on Computer Systems,8(2):85–110,May1990.

[4]P.Eugster,P.Kouznetsov,and R.Guerraoui.Reliable

Broadcast.Technical report,Swiss Federal Institute of Tech-nology,Lausanne,Jan.2001.

[5]B.Fenner,M.Handley,H.Holbrook,and I.Kouvelas.Proto-

col Independent Multicast-Sparse Mode(PIM-SM):Protocol Speci?cation.Internet Engineering Task Force,Nov.2000.

[6]V.Hadzilacos and S.Toueg.Fault-Tolerant Broadcasts and

Related Problems.In S.Mullender,Distributed Systems, chapter5,pages97–145.Addison-Wesley,2nd ed.,1993. [7]K.Lidl,J.Osborne,and J.Malcolm.Drinking from the?re-

hose:Multicast USENET news.In Proceedings of the Winter 1994USENIX Conference,pages33–45,jan1994.

[8]S.Paul,K.Sabnani,J.Lin,and S.Bhattacharyya.Reliable

Multicast Transport Protocol(RMTP).IEEE Journal on Se-lected Areas in Communications,15(3):407–421,Apr.1997.

词性转换

（一）译例解析 第一类：转译为动词 英语和汉英比较起来，汉语中动词用的比较多，例如，在He admires the President’s stated decision to fight for the job (他对总统声明为保住其职位而决心奋斗表示佩服)句中，英语的谓语动词只有admires一个词，其它用的是过去分词(stated)，动词派生的名词(decision)、不定式(to fight)和介词(for)。汉语没有词性变化，但可以几个动词连用，因此英语中不少词类，尤其是名词、介词、形容词、副词，在汉译时往往可以考虑转译为动词。例如： 1. He came to my home for help. 他来到我家，寻求帮助。 2. My admiration for him is growing more. 我越来越敬佩他。 3.Unfortunately, he was also a bit bossy, and he wasn’t a very good listener. 遗憾的是，他这个人有一点专横，不太善于听取别人的意见。 4.The operation of a machine needs some knowledge of it properties. 操作机器就需要懂得机器的某些性能。 5. Rockets have found application for the exploration of the universe. 火箭已经用来探索宇宙。 6.He is no smoker, but his father is a chain-smoker. 他不抽烟，但他爸爸却一只接一只不停地抽。 7.His mom is a good cook. 他妈妈做饭很好吃。 8.Talking with his so n, the old man was the forgiver of the young man’s past wrong doings. 老人和儿子谈话后，原谅了他过去干的坏事。 第二类：转译为名词 英语中很多由名词派生的动词，以及由名词转用的动词，在汉语中往往不易找到相应的动词，这时可将其转译成汉语名词。 1．She behaves as if she were a child. 她的举止像个孩子一样。 2．Glass is more transparent than plastic cloth. 玻璃的透明度比塑料布要好。 3. The film “ A Night t o Remember” impressed me deeply. 电影《冰海沉船》给我留下了深刻的印象。 4．This problem is no less important than that one. 这个问题的重要性不亚于那个问题。 5．Steinbeck defended the poor and the oppressed. 斯坦贝克替穷人说话，为被压迫者申辩 6. The new type of machine is shown schematically in Figure 1. 图一所示的是这种新型机器的简图。 7. Each of thee compounds boils at a different temperature. 这些化合物的沸点各不相同。 8.To them, he personified the absolute power. 在他们看来，他就是绝对权威的化身。 9.Stevenson was eloquent and elegant—but soft.

2017届高考英语二轮复习天天增分练(十五)

天天增分 （十五） 满分 48 分，实战模拟， 15 分钟拿下高考客观题满分 姓名： _ 班级： _____ Ⅰ.阅读理解 A Two years ago my grandmother was going to turn 75. My family discussed what the best way to celebrate was. Should we throw her a party? Should we take her on a trip? We remembered that she had touched so many people's lives ，and there were so many people for her to consider. Then someone got the idea that we should include everyone in the celebration by turning it into a tribute （ 献礼） to my grandmother. We secretly sent out letters to the people in grandmother's address book and asked them to send a letter with a memory that they had shared with her. People sent us letters with poems, stories and pictures. The deep feeling that was shared through the response surprised us. We compiled （ 编纂） these letters into a memory book and amazed her with it on the morning of her birthday. The unusual thing about my grandmother's friends was not the number that she had ， but the connection they shared. In many ways this book of friendship was the greatest achievement of my grandmother's life. I believe that developing true friendships is one of the most important things that anyone can do in one's lifetime. It is not a matter of the number of friends one has ，but the quality of the bonds. If one has had at least one true friendship before dying, then one can say he has lived a successful life. I have madem any friends and I believe I have begun to develop the same types of friendships my grandmother kept up over her lifetime. I only hope that I will be as successful as she has been. 1. How did the author's family celebrate the grandmother's birthday? A. They took her on a trip across the country. B. They gave her a memory book of friendship. C. They invited all her friends to her birthday party. D. They asked all her friends to send her cards. 2. When receiving her birthday gift, the author's grandmother probably felt A. disappointed and lonely B. sorry and sad C. surprised and pleased D. nervous and excited 4. According to the passage, the author probably agrees that . A. the more friends you have ，the better B. friends are more important than family C. understanding leads to greater success D. true friendship is very important to us 1．解析：细节理解题。根据第二段中的“ We secretly sent out letters to the people in grandmother's address book... ”和“We compiled （ 编纂）these letters into a memor book... ”可知， B 项正确。 3. The underlined word bonds ” in the last paragraph probably A. connections C. successes B. works D. celebrations

英语单词惯用法集锦解析

英语单词惯用法集锦 习惯接动词不定式的动词（V to inf) adore(vi极喜欢) dread （vt.不愿做，厌恶）plan 计划 afford（+to,vt有条件，能承担）endeavour （vt，竭力做到，试图或力图）prefer（vt.宁可；宁愿(选择)；更喜欢）agree 同意endure(忍受.cannot ~ to) prepare准备 aim (vi[口语]打算：) engage (vi.保证，担保；) presume(vt.冒昧；敢于[用于第一人称时为客套话]：) appear (vi.似乎；显得) essay(vt.尝试，试图) pretend(vt.自命；自称；敢于；妄为) apply （申请）expect（期望，希望）proceed（开始，着手，）arrange （vi.做安排，(事先)筹划）fail （vt.未做…；疏忽)promise(许诺，保证做 ask （要求）forget （vt. 忘记）purpose （vt.决心，打算） beg （vt.正式场合的礼貌用语]请(原谅)，请(允许)：I beg to differ.恕我不能赞同）guarantee（保证，担保）refuse（拒绝）bear 承受，忍受hate（[口语]不喜欢；不愿意；）regret （vt. 抱歉；遗憾）begin help （有助于，促进）remember（记住） bother （vi.通常用于否定句]麻烦，费心）hesitate（vi.犹豫；有疑虑，不愿）scheme（策划做）care （vt.想要；希望；欲望[后接不定式，常用于否定、疑问及条件句中]）hope （vt.希望，盼望，期待）seek（vt.谋求，图谋[后接不定式]） cease （停止; 不再(做某事)[正式] intend （打算；想要）seem（似乎，好像[后接不定式或从句]；觉得像是，以为[ choose （意愿；选定；决定）itch start开始claim （vt. 主张；断言；宣称） continue （继续）like 喜欢swear（vt.起誓保证；立誓要做(或遵守） dare （vt.敢，敢于，勇于，胆敢）long（vi.渴望；热望；极想） decline（vt.拒绝，拒不(做、进入、考虑等) manage（设法完成某事）threaten（vt.威胁，恐吓，恫吓）deign （屈尊做）mean（有意[不用进行时）trouble（vi.费心，费神；麻烦）demand（vi.要求，请求：）need （需要）try（设法做） deserve (应得) neglect (疏忽) undertake(承诺,答应，保证) desire （希望渴望）offer（表示愿意(做某事)，自愿；）venture（冒险(做某事））determine（vi.决心，决意，决定，）omit （疏忽，忘记）want 想要 die （誓死做）pine （渴望）wish （希望） 习惯接“疑问词+动词不定式”的动词（有时也包括VN wh-+to do) advise 建议explain 解释perceive 觉察，发觉 answer 答复find 得知，察觉persuade 说服，劝说；使某人相信 ask 询问，问forget 忘记phone 打电话 assure 保证guess 臆测，猜度pray 祈祷 beg 请求，恳求hear 小心聆听（法庭案件）promise 允诺 conceive 想象，设想imagine 以为，假象remember记得 consider 考虑，思考indicate 暗示remind 提醒，使想起 convince 使相信inform告知通知instruct告知，教导 see 看看，考虑，注意decide 解决，决定know 学得，得知 show 给人解释；示范；叙述；discover发现；知道learn 得知，获悉 signal以信号表示doubt 怀疑，不相信look 察看；检查；探明 strike 使想起；使突然想到；使认为suggest 提议，建议tell 显示，表明；看出，晓得；warn 警告，告诫think 想出；记忆，回忆；想出，明白wonder 纳闷，想知道 wire 打电报telegraph 打电报 习惯接动名词的动词（包括v+one’s/one+v+ing) acknowledge 认知，承认…之事实escape免除，避免omit疏忽，忽略 admit 承认，供认excuse 原谅overlook 放任，宽容，忽视adore （非正式）极为喜欢fancy 构想，幻想，想想postpone 延期，搁置 advise 劝告，建议finish完成prefer较喜欢 appreciate 为…表示感激(或感谢)forbid 不许，禁止prevent预防 avoid 逃避forget 忘记prohibit 禁止，妨碍

“的、地、得”用法分析及练习(后附答案)

“的、地、得”用法分析及练习（后附答案） 一、的、地、得用法分析： “的”后面跟的都是表示事物名称的词或词语，如：敬爱的总理、慈祥的老人、戴帽子的男孩、珍贵的教科书、鸟的天堂、伟大的祖国、有趣的情节、优雅的环境、可疑的情况、团结友爱的集体、他的妈妈、可爱的花儿、谁的橡皮、清清的河水...... “地”后面跟的都是表示动作的词或词语，如：高声地喊、愉快地唱、拼命地逃、疯狂地咒骂、严密地注视、一次又一次地握手、迅速地包围、沙沙地直响、斩钉截铁地说、从容不迫地申述、用力地踢、仔细地看、开心地笑笑......” “得”前面多数是表示动作的词或词语，少数是形容词；后面跟的都是形容事物状态的词或词语，表示怎么怎么样的，如：走得很快、踩得稀烂、疼得直叫唤、瘦得皮包骨头、红得发紫、气得双脚直跳、理解得十分深刻、乐得合不拢嘴、惊讶得目瞪口呆、大得很、扫得真干净、笑得多甜啊...... 二、的、地、得用法补充说明： 1、如果“de”的后面是“很、真、太”等这些词，十有八九用“得”。 2、有一种情况，如“他高兴得一蹦三尺高”这句话里，后面的“一蹦三尺高”虽然是表示动作的，但是它是来形容“高兴”的程度的，所以也应该用“得”。

三、的、地、得用法总结： 1、“的”前面的词语一般用来修饰、限制“的”后面的事物，说明“的”后面的事物怎么样。结构形式一般为：修饰、限制的词语+的+名词。 2、“地”前面的词语一般用来形容“地”后面的动作，说明“地”后面的动作怎么样。结构方式一般为：修饰、限制的词语+地+动词。 3、“得”后面的词语一般用来补充说明“得”前面的动作怎么样，结构形式一般为：动词（形容词）+得+补充、说明的词语。 四、的、地、得用法例句： 1. 蔚蓝色的海洋，波涛汹涌，无边无际。 2. 向日葵在微风中向我们轻轻地点头微笑。 3. 小明在海安儿童公园玩得很开心。 五、“的、地、得”的读音： “的、地、得”是现代汉语中高频度使用的三个结构助词，都起着连接作用；它们在普通话中都各自有着各自的不同的读音，但当他们附着在词，短语，句子的前面或后面，表示结构关系或某些附加意义的时候都读轻声“de”，没有语音上的区别。 但在书面语中有必要写成三个不同的字，这样可以区分他们在书面语用法上的不同。这样做的好处，就是可使书面语言精确化。

词性转换总结与归纳

词性转换总结与归纳 动词变名词 1.V+ment 结尾 1.announce — announcement 通知 2.equip 装备---equipment 装备,器材(不可数名词) equip---equi pp ed---equi pp ed---equi pp ing 3.settle— settlement 定居,安定 4.achieve---achievement 成就 5.amuse--amusement 娱乐 6.manage---management 经营管理 7.advertise--- advertisement广告 8.agree — agreement同意 9.disagree — disagreement 不同意 10.argue---argument 争吵 https://www.wendangku.net/doc/ba10590729.html,mit — commitment 奉献 commit---commi tt ed---commi tt ed---commi tt ing commit murder/suicide be committed to (doing) sth保证做某事 your commitment to work 你的敬业精神（投入/忠诚/奉献） 12.develop---development 发展 https://www.wendangku.net/doc/ba10590729.html,ern 统治— government 政府 14.judge---judgement/judgment

2. v+tion/sion结尾 15.describe — descri p tion 描写,描绘 16.erupt — eruption 爆发 17.affect — affection 喜爱、喜欢effect影响 unrequited affection单相思 18.satisf y— satisf action满意,满足 19.select — selection 挑选,选择 20.permi t— permi ss ion 允许 21.admi t— admi ss ion 承认,允许 22.invite — invitation 邀请,请帖 23.devote — devotion 献身,专注 24.apply — application 申请,申请书 25.produce — production 生产,产品 product 产品（常指工业制品）produce （农产品） 26.protect — protection 保护 https://www.wendangku.net/doc/ba10590729.html,cate-----education 教育 28.consider — consideration 考虑 29.attract — attraction 吸引力,吸引 30.instruct — instruction 指导, 介绍 31.discuss — discussion 讨论 32.appreciate — appreciation 感激,欣赏 33.recognize — recognition 认出

破解托福阅读理解长难句

破解托福阅读理解长难句 托福阅读理解难，难在长难句。如何破解长难句，是做好托福阅读理解的关键一环。破解长难句通常采用方法是找准主谓宾, 去除定状补，抓住长难句特点进行庖丁解牛是破解托福阅读理解长难句的又一制胜法宝。托福阅读理解长难句有以下六大类型： 1、主语、宾语拉长一般来讲，汉语主语比较简短。相比之下，英语中主语一旦拉长，就会增加读者的理解难度。解决方法：有效分解主谓成分，断开之后各个击破。 例如：（1）The deserts, which already occupy approximately a fourth of the Earth’s land surface, have in recent decades been increasing at an alarming pace. (OG Practice 2) （2）The destruction caused by the volcanic explosion of Mount St. Helens, in the northwestern United States, for example, pales in comparison to the destruction caused by humans. 2、分词短语打头，句子呈现三段（或三段以上）的长句式分词短语（包括现在分词和过去分词）做状语，这一句式比较常见，读者阅读时由于要区分短语和主语之间的逻辑关系，所以理解上有难度。并且分词短语本身往往由于掺入了插入语成分，理解上就更为困难。解决方法：理清主动和被动关系。一般来说现在分词与主语之间是主动关系；而过去分词与主语之间是被动关系。 例如：Working of a century-old schoolhouse in the village of Marland, Pennsylvania, the Conservancy’s Bud Smith is working with local people and business leaders to balance economic growth environmental protection. ☆注意：介词短语位于句首类似分词短语形式。 例如：In order for the structure to achieve the size and strength necessary to meet its purpose, architecture employs methods of support that, because they are based on physical laws, have changed little since people first discovered them—even while building materials have changed dramatically. (TPO 14) 3、多个谓语动词连用简单句之所以简单是因为成分单一便于读者理解。而托福阅读理解中，出题人为增加阅读难度，就会把几个谓语动词放在一个句子当中。但是无论局势如何变化，英语句子本身就犹如一棵大树，只能有一个主干起支撑作用，其他起辅助作用。解决方法：遇到多个谓语动词连用情况要分清主句谓语动词和从句的谓语动词。剔除细枝末节之后，句子也就好理解了。 例如：The same thing happens to this day, though on a smaller scale, wherever a sediment-laden river or stream emerges from a mountain valley onto relatively flat land, dropping its load as the current slows: the water usually spreads out fanwise, depositing the sediment in the form of a smooth, fan-shaped slope. (TPO Ground Water)

常见动词用法

1、keep ①keep + 形容词表示“保持” Please keep quite. 请保持安静。 ②keep + 宾语+ 形容词（或介词短语）表示“把……保持在某一状态” We must do everything we can to keep the air clean. 我们必须尽一切所能保持空气清洁。 ③keep sb doing sth 表示“让某人做某事” ——只能用现在分词作宾语补足语，不能用不定式。 He kept us waiting for two hours. 他让我们等了两个小时。 He kept us to wait for two hours. (错误) ④keep on doing sth和keep doing sth 表示“继续做某事，反复做某事”，可换用。 但keep on doing 更强调动作的反复性或做事人的决心。 He keeps on phoning me, but I don’t want to talk to him. Though he failed 3 times, he kept on trying. 他老是给我打电话，但我不想同他讲话。虽然他已失败了3次，但他仍继续干下去。 keep doing sth 经常用于静态动词。 He kept lying in bed all day long. 他整天都躺在床上。 ⑤keep …from doing sth 表示“阻止，使免于” He kept them from fishing in the lake. 他不让他们在那个湖里捕鱼。 2、may not / mustn’t / needn’t / wouldn’t ①may not be 是may be的否定式，意为“可能不是，也许不是” He may be there.他可能在那里。He may not be there.他可能不在那里。 ②must 意为“必须”，mustn’t 意为“千万不可，绝对不可” 所以Must we/I ……?的否定回答要用needn’t—意为“不必” -Must we get there before 11 o’clock？ -No, we needn’t. ③wouldn’t = would not 意为“不会，不愿” I wouldn’t say no. 3、do ①do表示“做”，做某事，常指某种不具体的活动；make表示“制作”，指做出某种具体的东西。

英语单词词性转换的基本规律

英语单词词性转换的基本规律 (1.动词(v.)→名词(n.) (a)词形不变,词性改变 例如:work, study, water, plant等可以用作动词(工作,学习,浇水,种植),也可以用作名词(工作,学习,水,植物). (b)一些动词在词尾加上-er或-or之后就变成了表示"某一类人"的名词 例如:work—worker, teach—teacher, sing—singer, jump—jumper, play—player, learn—learner, visit—visitor, invent—inventor, collect—collector等. 注意:1)以不发音的e结尾的动词,在词尾加-r. 例如:drive—driver, write—writer等. 2)以重读闭音节结尾,且末尾只有一个辅音字母的动词,应双写末尾的辅音字母,再加-er 例如：run—runner, win—winner, begin—beginner等. (c)在动词词尾加上-ment 变成名词 例如:achieve—achievement (成就) advertise—advertisement//advertising(广告) agree—agreement disgree—disagreement amuse—amusement (娱乐) improve—improvement(争吵) commit(奉献)—commitment develop—development (发展) depart—department (局,部) govern(统治)—government(政府) manage—management (管理) equip—equipment (装备) 有些单词比较特殊,需把动词后的e去掉再加ment. 例如:argue—argument(争论)

英语易混淆词汇知识整合辨析

一．表主动和被动的不同。 有些由动词词根派生而来的形容词与该形容词所修饰的词之间存在这逻辑上的主动或被动关系，其中以-able,-ible,和-ed后缀结尾的形容词往往含有被动意味，而以-ful,-ous，或-tive结尾的往往含有主动意味。 1.respectful与respectable(respecting,respective) respectful意思是“恭敬的，尊敬人的”，有主动性的“尊敬 他人”之意。例如：the children were well mannered and respectful toward grown-ups.孩子们彬彬有礼，尊敬大人。 Respectable是“可敬的，值得尊敬的”的意思，表示“还过 得去”，为一般性的赞语，含被动意味。例如： It’s not respectable to get drunk in the street.在大街上 喝的醉醺醺的是有失体统的。 Respecting常用作介词，表示“关于,,,,在，，，方面”之意。 例如： Problems respecting air pollution in cities should be considered seriously.关于城市空气污染问题应给与认真考虑。 Respective表示“各自的，分别的”的意思。 Go to your respective places.各就各位。 2.imaginative 与imaginable(imaginary) imaginative 意思是“富于想象力的”，和他所修饰的人之间 含有主动的关系。例如： teachers like imaginative children.老师喜欢富有想象力的

动词以及动词短语的用法(动词后加to do 还是doing)

一动词加-ing 的情况 consider, suggest/advise,look forward to, excuse/pardon admit,delay/put off,fancy avoid,miss,keep/keep on,practise deny,finish,enjoy/appreciate forbid,imagine,risk can't help,mind,allow/permit,escape 考虑建议盼原谅, 承认推迟没得想. 避免错过继续练, 否认完成就欣赏. 禁止想象才冒险, 不禁介意准逃亡. 如：建议：advise,suggest,冒险:risk,献身:devote oneself to 二动词后加doing 和加to do sth. 意思不一样的情况 ①remember doing指记住过去做过的事，remember to do指记住将来要做的事，表示“不要忘记”。 ②forget doing表示忘记过去做过的事，forget to do表示“没有想起做某事”。 ③mean doing表示“意味着做某事”，mean to do表示“打算做某事”。 ④regret doing表示对已做过的事感到后悔，regret to do表示对将要做的事表示遗憾。 ⑤stop doing表示“停止做某事”，stop to do是停止做正在做的事以便去做另外一件事，这里的to do不是stop的宾语而是stop的目的状语。 ⑥try doing表示“尝试做某事”，try to do表示“设法、试图做某事”。 ⑦go on doing表示继续做同一件事，go on to do表示做完一件事后，接下去做另外一件事。 三动词后加to do sth. afford负担得起agree同意appear似乎，显得arrange安排 ask问attempt企图beg请求begin开始 choose选择claim要求decide决定demand要求 desire愿望determine决定expect期望fail不能 forget忘记happen碰巧hate憎恨，厌恶hesitate犹豫 hope希望intend想要learn学习long渴望 love爱manage设法mean意欲，打算need需要 neglect忽视offer提供omit忽略，漏other扰乱；烦恼

词性转换73814

1 高中英语写作词汇词性转换 第一组： 1. ability (n. )能力；才能---able (a.)---unable (a..) —disability (n. ) —disabled (a.) be able to do…= be capable of doing 2. absence (n. ) 缺席；不 在 --- absent (a.) ---present (a..) – presence (n.) be absent from; be present at 3. absolute (a.) 绝对的；完全的---absolutely (adv.) 4. abundant (a.) 丰富的；充裕的 ---abundance (n.) be abundant in 5. academic (a.) 学院的，理论的 (n.) 大学教师 --- academy (n.) 学院 6. accept (v.) 接受---acceptable (a.)可接受的 - –refuse (opp.) 拒绝 7. access (n.) 通路，入门；（v.）接近；存取 ---accessible (a.) have access to（to为介词）可以使用；可以接触 8. accurate (a.) 正确的，精确的 --- accuracy (n.) –accurately (adv.)-- (opp.)inaccurate 不准确 9. achieve (v.) 取得，达到 --- achievement (n.) 10. acquire (v.) 获得，学到 --- acquisition (n.) 11. act (n. ) 法令，条例；（v.）表演；行动---action (n.) –active (a.) 主动的—inactive (a.) —actively (adv.) —activity (n. ) 活动 --passive (a.) 被动的 ---actor (n.) –actress (n.) 12. actual (a.) 实际的；现实的–actually (adv.) = as a matter of fact 13. adapt (v.) 使适应；改编 ---adaptation (n.) 适应，顺应；改编，改编本 ---adaptable (a.) 有适应能力的adapt oneself to 适应于; adapt…(for sth.) from sth. 改编14. add (v.) 添加，增加;补充说- --addition (n.) ---additional (a.) –additionally (adv.) add…to 将…(添)加到… add to 增加，加强 add up加起来 add up to共计达 in addition (to) 另外，除此之外 15. addict (n.) 成瘾的人 ---addicted (a.) –addition (n.) ---addictive (a.) 使成瘾的；上瘾的；易令人沉溺的 be addicted to （to为介词）对…上瘾 16. adjust (v.) 调整，使适应 ---adjustment (n.) --- adjustable (a.) 17. admire (v.) 钦佩，羡慕 ---admiration (n.) --- admirable (a.) 18. admit (v.) 承认，准许 ( admitted, admitted) --- admission (n.) 19. adopt (v.) 收养，采用 ---adoption (n.) --- adopted (a.) 20. advance (v.) 推进，促进；前进（n.）前进，提升—advanced (a.) 先进的;高等的 21. advantage (n.) 有点；好处 ---disadvantage (n.) take advantage of利用；欺骗，占…的便宜 22. adventure (n.) 冒险，奇遇 --- adventurous (a.) 23. advertise (v.) 为…做广告---advertisement (n.) 24. advise (v.) 建议，劝告----adviser (n.) 顾问，忠告者 --- advice (n.) a piece of advice 25. affect (v.) 影响 --- affection (n.) 影响，感情 = have an effect on 26. Africa (n.) 非洲 --- African (a.) (n.) 27. age (n.) 年纪 --- aged (a.)老年的—elderly (a.) 28. agree (v.) 同意--- agreement (n.) --- disagree (opp)—disagreement (n.) 29. agriculture (n.) 农业– agricultural (a.) 30. allow (v.) 准许，允许---allowance (n.) 允许；紧贴，补助

高频近义词辨析(1)

近/同义词辨析 1. reliant, reliable这一组形容词均源于rely, 含“依赖、依靠”之意 1) reliant依赖的，依靠的be reliant on sb/sth. 依赖 to be self- reliant 信赖自己，有自立能力，自强; self-confidence自信 self-reliant children / man有自立能力/自强的 be confident and self-reliant自信自强 To overcome reliant mind. 克服依赖心理。 The charity is completely reliant on public donations. 此慈善团体完全依靠公众捐款。 2) reliable可靠的,可信赖的trustworthy；稳妥的safe reliable information 确实可靠的消息 a reliable method 稳当的办法 be honest and reliable 笃实可靠 a reliable guarantee 确切的保证 Is your watch reliable? 你的表准吗? 2. fit，healthy，robust，sound，strong，sturdy，tough，vigorous， wholesome，well 健康的，强健的 1) healthy 指身体无病，也可指身心健全、正常的。 Swimming is a healthy pleasure.游泳是一种有益于健康的娱乐活动 2) sound侧重身体各部分或器官没有病，无任何缺陷，即健康。 His heart is basically sound. 他的心脏基本上健康。 3) well仅指没有疾病，但不一定很健康。 I don’t think he is really a well man.我认为他并不是真正健康的人。 4) strong既指体格健壮，又指体力或精神上的力量。 5) vigorous指人强健有力，精力充沛。 6) sturdy侧重于结实的体格。 7) tough着重指人的体格健壮。 8) wholesome多指能给人留下身体健康、思想健全或品德良好等印象。 9) fit既可指健壮又可指健全无病。 10) robust强调身体强健。 3. abandon，desert，forsake，quit，leave，give up抛弃，放弃 1) abandon强调永远或完全放弃或抛弃人或事，尤指对之负有责任或义务者，放 弃一个项目或计划, 如abandon principles / research / sb.

pretend三种易混淆不定式的用法

pretend三种易混淆不定式的用法 今天给大家带来了pretend三种易混淆不定式的用法，我们一起来学习吧，下面就和大家分享，来欣赏一下吧。 pretend三种易混淆不定式的用法 1. pretend to do sth .这个短语的意思是假装(将要)去做什么事情，适用于将来时态动作将来假装要去做但不一定去做的状态。 举例: If youpretend to know what you dont know，youll only make afool of yourself.不懂装懂就会闹笑话。(suri的回答) Child pretend to be mother and father in kindergarten.孩子在幼稚园里面假扮父亲和母亲(表将来)(JasoOon的回答) 2. pretend to have done sth .这个短语的意思是假装已经做过了某事，强调事情的一个完成的状态，侧重于假装的事情已经做好了。 举例:I pretend tohave seen nothing，but I cant.我假装自己没有看到任何东西，但是我做不到(侧重于一个完成时态，已经试图去假装没有看到)

she pretended to have finished the homeworkwhen she went out and played.当她出门玩的时候她假装自己已经完成了家庭作业。(假装做作业这个动作已经在出门玩之前做完了)(JasoOon的回答)以及怀陌的回答:When the teacher came in，he pretended to havefinished the homework.当老师进来的时候他假装自己已经完成家庭作业了，两者有异曲同工之妙。 3. pretendtobe doing sth 这个短语的意思是假装正在做某事，强调动作的一个进行时态。 举例:They pretend to be reading books when the teacher sneakingly stands at the back door.当老师偷偷地站在后门的时候他们假装正在读书(读书与老师站在后门都是过去进行时 态)(JasoOon的回答) Asmanypeople do，youoftenpretend to be doingwork when actuallyyou arejust wasting time online.像很多人一样，你经常假装正在工作，其实是在上网。 群主补充：昨天和今天已提交作业的同学，做得都很好，全部授予小红花。希望你们再接再厉，不要松懈哟。所以下周一出题者为所有已提交作业的同学或者你们选出的代表。