An approach to a methodology and implementation of a Network Monitoring Analysis
Tool
Elias Aravantinos Dr. Christos Bouras
Dr. Petros Ganos Computer Engineering and Informatics Department
Aristidis Ilias University of Patras Research Academic Computer
Technology Institute
GR-26500, Patras, Greece
Riga Feraiou 61Research Academic Computer Technology Institute
GR- 26221 Patras, Greece E-mail: bouras@cti.gr E-mail: eliasara@cti.gr
E-mail: ganos@cti.gr
E-mail: ilsadis@cti.gr
KEYWORDS
Network monitoring, IP Based Networks, SNMP, Real Time System, Network resources, Databases.
ABSTRACT
In this paper we describe the design and implementation of a network monitoring analysis tool. The network resources were configured to support SNMP under various operating systems, thresholds definitions and events. The data were collected by a network monitoring system and handled according to the values of the variables or the event type. A real- time report to a database was established via adaptive scripts. A Network Node Manager (NNM) will inform an external, remote database about the network status by sending real time data containing alerts and events. The whole tool is called Network Management Analysis Tool (NMAT).
INTRODUCTION
A fundamental division among network monitoring systems is related to whether the monitoring is done off-line (while only test messages are flowing), on-line (while user traffic is on the network) or both. Techniques for monitoring networks, when they are out of service, involve the generation of test sequences, monitoring and analysis of the results, to determine measurable levels of performance and protocol conformance. Such tests may be implemented in a test bench environment, manufacturing floor, network operations centre or during an outage in an operational network.
The process of on-line monitoring and analysis may be performed on a continuous basis, scheduled at various times of day or invoked only when circumstances (such as load changes or quality concerns) demand a closer scrutiny. Continuous monitoring (intrusive testing) is the most costly alternative in terms of resources and time, but is also the most beneficial from a user’s viewpoint.
Monitoring may be divided into two categories according to the layers involved: Monitoring and analysis of the physical and network layers referring to the OSI model
Monitoring and analysis of the hardware platforms and operating systems
A very large number of access points (network nodes) may be required and they would usually be dedicated to the monitoring system. The amount of data that needs to be collected and transported to a local/remote database may also be very high. The benefit, however, is that proactive management and dynamic prediction may be a reality. Network performance measurement is an extremely broad area and we can only briefly mention some of the more relevant work. (Carter and Crovella 1996) present two tools “bprobe” and “cprobe” to measure the bottleneck link speed and competing traffic respectively, on a path using ICMP ECHO packets. Since these tools do not use TCP, they are not able to capture any TCP related effects that an application might experience. Van Jacobson proposes the tool “pathchar” (Jacobson 1997) that estimates bandwidth on all hops of a path and hence can take a very long time. It also requires root access, making it less desirable for grid environments. The tool “Treno” (Mathis amd Mahdavi 1996) emulates an idealized TCP, which makes the measurements independent of host-specific TCP implementations but not representative of what applications would experience.“Treno” also needs root access. Topology-d (Obraczka 1998) uses “ping” and “netperf” to make measurements between all pairs within a group and then computes a minimum-cost logical topology. The Network Weather Service (NWS) (Wolski 1998) uses TCP to send small, fixed-size probes measured in kilobytes, with a frequency that is tuneable, but typically ranges from tens of seconds to several minutes. Performance measurement systems, such as the National Internet Measurement Infrastructure (NIMI) project (Paxson 1998) are designed for arbitrary Internet hosts. The Cooperative Association complements this work for Internet Data Analysis. The goal is to develop metrics and tools for analyzing traffic across the Internet. Additionally, the data from such tools intend to be used by “higher-level” systems. Lowecamp (Lowecamp 1998) also provides an API whereby applications can pose flow and topology-based queries.
AppLeS (Application Level Scheduler) (Su 1998) uses information from the NWS to schedule distributed resource-intensive applications.
Finally other components will use “Gloperf” information in similar resource discovery and allocation functions. We note (Lee 1999) that “Gloperf” is designed to enhance portability. It makes end-to-end TCP measurements. Storing “Gloperf”data in a directory service provides data discovery and access. However, there are missing some measurements to allow applications to probe network resources and getting fresh data.
The purpose of this paper is to approach a methodology and implementation of a Network Monitoring Analysis Tool (NMAT) during the operation of an intranet network containing hundreds of terminals established in enterprise buildings, airplanes, ships etc. The aim is to collect automatically and in real-time values of monitoring data in a storing schema, a database for instance and finally make decisions at a human level.
This paper is organized as follows: First the tool design is presented in main principles. The next section describes the proposed phases of functionalities, containing techniques, specifications and results of network monitoring. Finally future work and concluding remarks are provided. ARCHITECTURE
The Network Management System (NMS) used, is HP Open View Network Node Manager (NNM). It is a commercial, well-known and reliable platform that implements the Simple Network Management Protocol (SNMP). The NNM collects by polling in tuneable time intervals the SNMP data from the managed network devices of the intranet and automatically sends the required values of data to a database.
The tables are filled by these values, which could be exported in various metadata formats in order to realize a data statistical analysis. Then the analysis will be assessed and will also be the input to form the predictions of network behaviour and performance. The last step concerns decisions and reconfiguration processes.
The whole process is described as following:
The NMS is fully configured.
The general database is created (installation, table creation).
The network resources are configured to support SNMP (agent) under Linux and Windows operating system.
Thresholds are defined and events are configured at the NMS.
Data are collected by the NMS and handled according to the values of the variables or the event type.
Real-time report is imported to the database via a script language.
The events are classified by an appropriate value for later data manipulation and analysis.
Prediction and decision analysis about the network components performance are made.
The tested network is an Intranet, supported by a Fast Ethernet network with bus topology. The Fast Ethernet network cable is Unshielded Twisted Pair category 5 (UTP cat 5). The tested network consists of a certain number of terminals, servers, Local Area Network (LAN) router and LAN switches. Each LAN segment is located in a different physical place in the Intranet. The LAN switches are Fast Ethernet network devices that provide each sender/receiver pair with full 100 Mbps capacity. Each port on the switch gives full bandwidth to a single server or client station. Furthermore, there is a switch-router connected to the Internet to serve the Intranet. The terminals and the servers are connected to different LAN segments. It was supposed that the tested terminals are located in different physical LANs instead of VLANs. All the terminals are connected to different switches instead of VLANs of the same switch.
Figures 1: System Architecture
Figure 1 describes the tested system architecture. The network monitoring process was based on the SNMP, managing all devices, like servers, switches and PCs. The Remote Database was hosted on a LINUX Server and NMS established a one way communication link, each time that appeared the need of sending data to the database.
The NMS server supported Windows 2000 Server operating system and the terminals Windows 2000 and LINUX.
An SNMP agent resides in each terminal machine in order to be remotely managed by NMS. The managed devices run software (SNMP agent) that enables them to send alerts (traps) when they recognize problems (for example, SNMP link failure). Upon receiving these alerts, management entities are programmed to react by executing one, several or a group of actions (referring to scripts) including event logging, changes in device configuration and system shutdown.
The management entities poll the end stations to check the values of certain variables in Management Information Bases (MIB). A MIB is a SNMP structure that describes the particular device being monitored. Polling can be automatic or user-initiated and agents in the managed devices respond to all polls. Agents are software modules that first compile information about the managed devices in which they reside, then store this information in a management database and finally provide it (proactively or reactively) to management entities within NNM via SNMP.
IMPLEMENTATION ISSUES
In our tested network an external database is used instead of an internal NNM database. The reason is to create applicable tables with different relations compared to the compact, complicated and inflexible NNM’s database model. As a result we propose a scheme, which could appear database
platform independency. Moreover the use of an external database provides interoperability, because it is possible to comply with several software tools, like JAVA-based platforms or other experimental tools developed for similar purposes.
The NNM collects by polling in tuneable time intervals the SNMP data from the managed network devices of the intranet and automatically sends the required values of data to a database.
Any corporate network has two types of end users: the organization’s typical employee and the network operator who is responsible for maintaining the end-to-end service. The organization’s typical end users and the success of the business applications that live on the network are the ultimate beneficiaries of all the various network monitoring activities that are performed within the network. However, the end users are least aware of the infrastructure behind the service. They simply need a quality of service that meets their business needs and achieves certain invisibility. It is the goal of achieving this “invisibility” of consistently good service that makes the added cost of monitoring and analysis tools justifiable.
The MIBs are used to monitor and access network and system information variables. They are either vendor MIBs or experimental MIBs. The vendor MIBs concern mainly the network switches while the experimental ones are used to monitor special parameters like the hostmib.mib which monitors system information (for example CPU load).
The SNMP polling interval was generally configured at 60 seconds. The NMS polls each critical variable every 10 seconds receiving traps and events. This polling interval concerns mostly the servers and the switches of the Intranet. The rest of the nodes are controlled from NNM every 1 minute.
The Network Management Analysis Tool (NMAT) reports all events in the external database with the help of two script-files (Brown 1999) (Events.ovpl and Events_value.ovpl) developed in Perl scripting language. These files are useful for the automatic actions, in order to insert the network variables values polled from NNM into the Database. Both scripts are running with the help of Open View Perl installed with NNM and configured with the help of the Event Configuration module of NNM.
The main problem we had to solve was the handle of consequential events, which are events that occurred exactly the same time. After making several experiments, we noticed that there was a unique variable generated by every event. This variable helped to distinguish the consequent events as discrete instants and store them into a remote database.
The events are separated in two categories:
Events with value
Events without a measured value
Events with value
The configuration of the events that do contain value should include the following automatic action command: OVHIDESHELL cmd.exe /c "Events_value.ovpl $s $ar $N $3 $8"Through this command the Perl file is called and rest of the symbols present the specific event variables based on NNM’s manual.
#!/opt/OV/bin/Perl/bin/perl
my $param_1 = $ARGV[0];
my $param_2 = $ARGV[1];
my $param_3 = $ARGV[2];
my $param_5 = $ARGV[4];
my $param_4 = 100;
my $param_7 = 100;
my
$SUMMARY_TEXT="tmp_event_".$param_5.".sql"
;
open(SUMMARY,">$SUMMARY_TEXT");
$command1 = "insert into nms(severity,source,event,value)
values ('$param_1', '$param_2', '$param_3', '$param_4');";
printf SUMMARY ("$command1");
close(SUMMARY);
$command2
=
"psql.exe -h 150.140.21.70 -f $SUMMARY_TEXT system nms";
$command3 = "del $SUMMARY_TEXT";
system($command2);
system($command3);
exit 0;
Figures 2: Listing of Events_value.ovpl
Events without a measured value
The configuration of the events that do not contain value should include the following automatic action command: OVHIDESHELL cmd.exe /c "events.ovpl $s $ar $N 100 $3"
#!/opt/OV/bin/Perl/bin/perl
my $param_1 = $ARGV[0];
my $param_2 = $ARGV[1];
my $param_3 = $ARGV[2];
my $param_4 = $ARGV[3];
my $param_5 = $ARGV[4];
my
$SUMMARY_TEXT="tmp_event_".$param_4.".sql"
;
open(SUMMARY,">$SUMMARY_TEXT");
$command1 = "insert into nms(severity,source,event,value)
values ('$param_1', '$param_2', '$param_3', '$param_5');";
printf SUMMARY ("$command1");
close(SUMMARY);
$command2
=
"psql.exe -h 150.140.21.70 -f $SUMMARY_TEXT system nms";
$command3 = "del $SUMMARY_TEXT";
system($command2);
system($command3);
exit 0;
Figures 3: Listing of Events.ovpl
The scripts use some NNM variables that are generated with the event, open a connection with the Postgres database using
the “psql” client and insert data to the NMS table. By the time of an event, a trap occurs, an automatic action occurs,the appropriate script is executed reporting the status of the nodes (up or down) and certain data are imported to the NMS table of the database. The data could be event description,severity of the event (normal, major) and special values such as CPU load and temperature. When the event does not contain value like node status events, Events.ovpl is running.When the event does contain value like CPU load Events_value.ovpl is running. Additionally, PSQL client helps to connect to the database by using the IP of the LINUX server instead of the DNS to avoid a possible failure of DNS server.
Then, the events should be activated (this is an easy procedure). Once the configuration is created, everything is included in the trapd.conf file. The events and the specifications are:
Table 1: Events Configuration
Specifications
Name of the event Polled by the SNMP manager every 60 sec OV_NODE_UP OV_NODE_DOWN
Checked on all machines every 60sec
OVER the limit in case load > 70%
REARM the limit in case load < 60%OV_CPU_OVER OV_CPU_REARM
OVER if T° > 41°C REARM if T° <40°C
OV_TEMP_OVER OV_TEMP_REARM Link is DOWN if connection of cable is removed
SNMP_LINK_UP SNMP_LINK_DOWN
The thresholds concern the control of CPU load (OVER in case load > 70%, REARM in case load < 60%) of all monitored nodes and also the control of CPU temperature (OVER if T° > 41°C, REARM if T° <= 40°C) at terminals side. When the threshold value is exceeded, NNM reports to the database through an automatic action the new value.Everything is included in the SnmpCol.conf file.
In order to keep NMS CPU level in low level, a task is added in Windows task manager that restarts the “SNMPcollect”procedure every 5 minutes, running cpuNMS.bat file. This task is necessary because the CPU is overloaded due to the short SNMP polling intervals. The suggested regular NNM polling interval from HP is 5 min.
The external database could be Postgres, MySQL etc. The database management system selected was a Postgres, hosted on a LINUX server. The configuration of DBMS concerned only trusted “users”, specifically only the IPs of NMS server represented a trusted machine, which could communicate with the appropriate port of the LINUX server and with the
database. The above option was selected to ensure the security of database and its data.
NNM is able to send and store data to the central database.The interface uses the Postgres client to connect to the database and some SQL commands to insert data into the
NMS table. All the modules of this interface are stored in a temporary folder of hard disk.
Table 2: NMS Table Architecture
ID Event Identifier Event Short description
Source The IP Address of the Node Severity Event Importance
Value
Variable value of measured events
A different script is executed according to the type of the event that occurs, in order to store CPU load, node status,temperature etc. into the database table. The script is embedded in the Events configuration module and exports several variable values of NNM. All actions are automated and occur in real time. The following table shows a sample record in the NMS table:
Figures 4: Sample record
The first field is an auto-counter, next the severity importance, next the source indicating the IP address, next a field that contains the local date-time and last a field that contains a short description of the event. Postgres generates automatically the field date-time. In the field “Value” is stored data from measured events like temperature and CPU load.
Other variables that NNM monitors and measures are: Interface Utilization of TCP/IP port Interface errors Bandwidth
Hardware fails of the monitored devices CPU load
Temperature value of the terminal Node status
Thresholds to critical values SNMP traffic, etc.
Table 3 presents some samples of database records during a certain time period of about three months. It is obvious that the desired network and system variables were collected according to the scenario. The Event with ID 11298 that was stored in the table referred to a very important, major situation affecting the CPU of the machine with IP 150.140.21.63. The CPU load of the machine at the specific date time was over the specific limit, about 80%. Then the Event with ID 11302 refers to a situation where the CPU operates normally according to pre-configured limits. The network administrator checks the table and finds out which are the ‘weak’ points of the network. Then he makes decisions about network and system improvements, in order to avoid some similar situations or combination that already occurred. The main and critical issue is to take decisions about the whole information system containing network expansion, bandwidth issues, devices with hardware and software specifications etc. concluding sometimes to a brand new system ‘refresh’, after taking into account the different parameters.
Table 3: Sample of experimental results
ID EVENT SOURCE SEVERITY VALUE DATE AND TIME
8356OV_Node_Up150.140.21.39Normal 27/08/2002 01:34:32
11298OV_Cpu_Over150.140.21.63Major8017/10/2002 01:41:15
11299OV_Node_Down150.140.21.30Normal 17/10/2002 01:41:43
11302OV_Cpu_Rearm150.140.21.63Normal2017/10/2002 01:51:01
13825SNMP_Link_Down150.140.21.38Major 20/11/2002 10:35:08
FUTURE WORK
The future work concerns the extension of the monitored SNMP variables in order to cover all the possible network events. The final expectation is to focus on the real network status, show its weak points and improve the network performance.
The prediction and decision analysis of the data collection could become another issue of our future work. We intend to select some standard models related to data analysis and do further work close to this problem. This work will enhance the decision analysis and prediction parts of our tool. Another future goal is to apply this tool to several networks and assess the variables based on their criticality.
CONCLUSIONS
For network operators, network monitoring and analysis provides the means to become proactive (i.e. to detect faults prior to receiving a user’s complaint). It also allows them to manage service level contracts, to be assured of day-to-day operations and to validate system changes.
The result of our work is a methodology and an implementation of a network monitoring analysis tool, which could improve the network performance. This could be achieved via the prediction and decision analysis of the data collection.
REFERENCES
Brown, M. C., 1999, “The Complete Reference Perl”, Osborne/McGraw-Hill.
Carter, R. and M. Crovella. 1996. “Dynamic server selection using bandwidth probing in wide-area networks”, Technical report, Boston U., TR-96-007.
Carter, R. and M. Crovella. 1996. “Measuring bottleneck link speed in packet-switched networks”, Technical report, Boston U., TR-96-006.
Jacobson, V. 1997. “A tool to infer characteristics of Internet paths”, Technical report, Lawrence Berkeley Lab.
Lee, C. A., J. Stepanek, R. Wolski, C. Kesselman and I. Foster, November 1999, “A Network Performance Tool for Grid Environments”, Technical Paper presented at Super Computing ‘99, Portland, Oregon, USA.
Lowecamp, B. et al., August 1998, “A resource query interface for network-aware applications”, 7th IEEE Symposium on High Performance Distributed Computing, pages 189–196.Mathis, M. and J. Mahdavi, 1996, “Diagnosing Internet congestion with a transport layer performance tool”, Proc. INET ’96. Obraczka, K. and G. Gheorghiu. August 1998. “The performance of
a service for network-aware applications”, 2nd Sigmetrics
conference on parallel and distributed tools.
Paxson, V., J. Mahdavi, A. Adams, and M. Mathis. August 1998.
“An architecture for large-scale internet measurement”, IEEE Communications, 36(8): 48–54.
Su, A., F. Berman, R. Wolski and M. M. Strout, November 1998,“Using apples to schedule a distributed visualization on the computational grid”.
Wolski, R., N. Spring, and H. Hayes. 1998. “The network weather service: A distributed resource performance forecasting service for meta-computing”, Future Generation Computing Systems.
--- to do与doing的区另U 一般情况下，to do 是一般将来式,是打算去做什么（未做）;doing是现在进行式,是现在正在做什么，或（此事已做过或已发生、正做） like to do 和like doing 的用法有什么区别 简单的记忆方法。当表示喜欢，用like doing ，如：He likes cooking in his house. She likes singing. 表示爱好。 当表示想要，欲做某事（但还没进行）用like to do ，例如：He likes to cook in his house.- 他想在自己家做饭吃。 She likes to stay with us.- 她想和我们带一块儿。（但还没进行） 2 forget doin g/to do forget to do 忘记要去做某事。（未做） forget doing 忘记做过某事。（已做） The light in the office is stil on. He forgot to turn it off. 办公室的灯还在亮着，它忘记关了。（没有做关灯的动作） He forgot turning the light off. 他忘记他已经关了灯了。（已做过关灯的动作） Don't forget to come tomorrow. 别忘了明天来。（to come动作未做） 3 remember doin g/to do remember to do 记得去做某事（未做） remember doing 记得做过某事（已做） Remember to go to the post office after school. 记着放学后去趟邮局。 Don't you remember see ing the man before? 你不记得以前见过那个人吗？ 感官动词 see, watch, observe, notice, look at, hear, listen to, smell, taste, feel +doing表示动作的连续性，进行性 I saw him working in the garden yesterday. （强调”我见他正干活”这个动作） 昨天我见他正在花园里干活。
浅谈现代汉语和英语 在语音、语汇、语法方面特点的比较 2012级学科英语：王忠文随着世界性汉语热的影响，学习汉语的人越来越多，而在交际交流中，中、英文的交叉使用现象也越来越多，那么对现代汉语和现代英语之间在语音、语汇、语法方面做一个比较也显得尤其必要。 一、语音即语言的声音，它是指人类通过发音器官发出来的，具有一定意义的，用来进行社会交际的声音。语音是语言的物质外壳，人们是通过语音来感知语言的存在的。 现代汉语语音属于汉藏语系，和印欧语相比有许多显著的特点。在语音方面，现代汉语的总体特点是音节界限分明，乐音较多，加上声调的高低变化和语调的抑扬顿挫，现代汉语语音的音乐性较强，具体表现为： 1.汉语音节中没有复辅音，都是一个辅音，而且不会出现在音节开始，即使是在词尾也只限辅音[n]和辅音群[ng]，如[an]案和[hong]红。而英语中的辅音或辅音群是常见的，他们可以出现在词首或词尾，常常两三个辅音放在一起。 2. 元音占优势。现代汉语中可以没有辅音，但元音是必不可少的成分，并且是乐音，听起来比较悦耳。而英语的音节中最少须有一个元音，最多可以有8个元音。闭音节多，开音节少；音节中以辅音占优势，而且在元音的前或后都可以有辅音
群出现。 3.音节有声调。语音中超音段表达的主要手段是音高、音强和音长。而音高的使用又最为广泛，它的语音物质表现形式为旋律（melody）,当旋律以单个的音节或词作为基本载体时，被称作声调（tone）；当旋律以短语和句子作为基本载体时，被称作语调（intonation）。声调或语调是体现音高变化的单位。 汉语是声调语言，音高变化分布在字，即音节上，汉语中每个音节都是由声，调和韵组成，一个汉字即一个音节。不仅有区别意义的作用，还可以使汉语音节界限分明，又富于高低升降的变化，从而形成了现代汉语语音音乐性强的特殊风格。而英语是语调语言，音高变化分布在短语和句子上。英语句子语调由句子的最后一个重读音节即语调核心表现出来，句子中的语调由语调群表现出来。而语调群分为调头、调体和调尾组成。英语中语调的特点如下：一般地说，低调冠，高调头，调身用平调且依次渐变滑动降低。语调核心接受句子语调，有调的变化，或升调或降调。由此，汉、英两种语言的语音对比，主要是音高变化的对比。 二、语汇是语言中语词的总汇。它是语言的建筑材料，是语言的构成要素之一。语汇的含义有广义和狭义之分：广义的语汇，指的地语言中的语和词（简称语词）的总汇；狭义的语汇，指的是语言中的语的总汇。我们这里所说的语汇，其含义是广义的，即指的是语言中语和词的总汇，与传统的包括语和
英语中常见可分可合单词的用法区别 英语语中常见可分可合单词的用法区别在英语中，有时两个部分合在一起构成一个词，表达的是一种意思；两个部分分开，各成一个词时，表达另一种意思，有另外的用法。现将中学英语课本中出现的这类“分合有别、用法各异”的 词语归纳分析如下： 1. away 与 a way away 用作副词时，意为“远”、“离开”、“远离”，通常作 be 的表语，也可以放在动词或副词之后，与动词一起构成短语动词；a way 是名词短语，是“一条 路”、“一个方法”之意。例如： Who will look after these children when I'm away? 我走了之后，谁来照看这些孩子？ The village is about eight miles away from here. 那个村子离这儿大约有八英里远。 She thought hard， and she found a way at last. 她苦思冥想，终于找到了一个方法。 2. already 与 all ready already 是副词，意为“已经”，常与完成时态连用；all ready 是形容词短语，常用作表语，意为“一切就绪”、“都准备好了”。例如： We have already finished our work. 我们已经做完了我们的工作。
We are all ready for the trip. Let's go. 我们都做好了旅行的准备。那么咱们就 上路吧。 3. awhile 与 a while awhile 是副词，意思是“一会儿”、“片刻”，其前不可加介词或副词； a whil e 是名词短语，意思是“一会儿”(与 awhile 同义)，a while 前面可以加介词in，af ter或for，也可加副词quite。例如： Wait awhile. [= Wait (for) a while. ] 等一会儿。 The postman came quite a while ago. 邮递员好一会儿前来过。 After a while, they quarrelled again. 过了一会儿，他们又吵了起来。 4. asleep 与 a sleep asleep 是形容词，意为“睡着的”，只能跟在fall 和be 后面作表语，或置于f ind 后面作宾语补足语；a sleep 是名词短语，意为“一段睡眠”，与have 连用， 构成 have a sleep (睡一会)。例如： Don't make any noise. The baby is asleep. 别闹，小孩睡着了。 He was just falling asleep when there was a loud knock at the door. 他刚要入睡，这时传来很响的敲门声。 I had a sound sleep last night. 昨晚我睡得很香。 5. anyway 与 any way
A和An的意思其实是一样的，都表示一个或单数的意思。当修饰的名字开头是元音（特别注意：是发音是元音而不是字母是元音字母啊）时，用an，例如an interesting book，而a则用于辅音开头的单词，例如a happy ending 名词的第一个发音为元音音标时，用an eg: an eye, an ear 名词的第一个发音为辅音音标时，用a eg: a hand 英语字母u前面的不定冠词a，an 的确定让学生头痛，怎么掌握呢？很简单，U 发字母表音的单词前是a，不发字母表音的单词前是an： a university 一所大学 a useful book 一本有用的书 an umbrella 一把雨伞 an underground train 一部地铁列车 冠词一般位于所限定的名词前，用来署名名词所指的人或事物。冠词有不定冠词和定冠词两种。不定冠词有两个形式，即a和an。a用在以辅音音素开头的词前，如a book; an用在以元音音素开头的字母前，如an apple. a或an与可数名词单数连用，泛指某类人或某物中的一个。the既可以用在可数名词前，也可以用在不可数名词前，表示某个或某些特定的人或事物，也可以指上文提到过的人或事物。 the/an/a用法及区别： 冠词在英文中属于虚词，在句子里主要是对名词起限定作用。这类词的数量很小，但是其活动性却很大，很大。在英文中，a / an并不等于one 。 例：He only has one dictionary.他只有一本词典。（正确） He only has a dictionary.他只有一本词典。（错误） One minute is enough for me to do it . 我做这件事一分钟足够了。（正确） A minute is enough for me to do it . 我做这件事一分钟足够了（错误） 1）不定冠词在句子中最大的语法功能是：用在可数名词的单数形式前表示泛指--表明一类人或事物区别于它类。 例：I am a Chinese. 我是(一个)中国人。 This is a book. 这是（一本）书。 2）为了读音的方便，在以元音音素开头的可数 名词的单数名词前用an而不用a。当我们使用an时，条件有三：①这个名词的读音必须是以元音音素开头--即它的音标的第一个音素是元音，而不是说它是以元音字母开头。②它必须是个可数名词。③它还必须是个单数名词。我们常常见到这类用法： a university 一所大学an hour 一个小时 an orange 一只桔子an engineer 一位工程师 an ordinary man一个普通人 an honest person一位诚实的人
加to do 的动词 attempt企图enable能 够 neglect忽视afford负担得 起 demand要求long渴 望 arrange安排destine注 定 mean意欲，打算begin开 始 expect期望omit忽略，漏 appear似乎，显得determine决定manage设 法cease停止 hate憎恨，厌恶pretend假装 ask问dread害 怕 need需要agree同 意
desire愿望love 爱 swear宣誓volunteer志愿 wish希望bear承 受 endeavor努力offer提 供 beg请求fail不 能 plan计划 bother扰乱；烦恼forget忘 记 prefer喜欢，宁愿care关心，喜欢happen碰 巧prepare准 备decide决 定learn学 习 regret抱歉，遗憾choose选择hesitate犹 豫profess表明
claim要求hope希 望 promise承诺，允许start开始undertake承 接want想要 consent同意，赞同intend想要refuse拒 绝decide决定 learn学习vow起contrive设法，图谋incline有…倾向propose提议seek 找，寻觅 try试图 2)下面的动词要求不定式做宾补：动词+宾语+动词不定式 ask要求，邀请get请，得 到 prompt促使allow允 许 forbid禁止prefer喜欢，宁愿announce宣 布force强
迫 press迫使bride 收 买 inspire鼓舞request请求 assist协助hate憎 恶 pronounce断定，表示advise 劝告exhort告诫，勉 励pray请求 authorize授权，委托help帮 助recommend劝告，推荐bear容 忍implore恳 求remind提醒 beg请求induce引 诱 report报告compel强 迫 invite吸引，邀请，summon传 唤command命 令intend想要，企
2017年高考英语易错词汇：some times用法区别 sometime,sometimes,some time,some times用法区别 sometime (1)副词：某个时候。可指过去或将来的某个时候。如： ① Would you come to the cinema with me， sometime? 赏脸找个时间和我看场电影吗?(将来) ② We‘ll take our holiday sometime in August. 我们会在八月找个时间度假。(将来) ③ I bought this sometime last summer. 这是我在上个夏天买的。(过去) ④ This mansion was built sometime around 1980. 这栋大厦是1980年左右建的。(过去) (2)形容词：前，过去的。意思类似former. ① This is our sometime general manager.这是我们的前任总经理。 ② Our sometime classmate， Rose， is now a pop star. 我们以前的同学罗斯，现在是明星了。 sometimes 副词：有时候。顺便提一下，at times也是“有时”的意思。 ① Every man is a fool sometimes， and none at all times. 每个人有时都会犯傻，但没有谁会一直都傻。 ② I sometimes have letters from him. 有时我会收到他的来信。 some time 某段时间。常与for连用。
常见的“to do”与“doing”现象 有些动词后既可接to do，也可接doing，它们后接to do与doing在意思上有时有较大的差别。因为它们也是中考的常考点之一，因而我们应该搞清楚它们的区别。 1. stop to do/stop doing sth。 解析：stop to do sth.意为“停下来（正在做的事）去做（另外的）某事”，to do sth.在句中作目的状语。而stop doing sth.意为“停止做（正在做的）某事”。如Mary stopped to speak to me.玛丽停下（手头的工作）来跟我讲话。 When the teacher came in. the students stopped talking.老师进来时，学生们停止讲话。 2. remember to do/remember doing sth 解析：remember to do sth.意为“记住要去做某事”（还没有做）。而remember doing sth.意为“记得（已经）做过某事”如： Please remember to send the letter for me.请记住为我发这封信。 I don’t remember eating such food somewhere.我不记得在哪里吃过这种食物 3. forget to do/forget doing sth 解析：forget to do sth.意为“忘记做某事”（动作还没有发生）。而forget doing sth.意为“忘记做过某事”（动作已发生）。如： Don’t forget to bring your photo here.别忘了把你的相片带来。 I have forgotten giving the book to him.我忘记我已把书给了他。 4. go on to do/go on doing sth 解析：go on to do sth.意为“做完一件事，接着做另外一件事”，两件事之间有可能有某种联系。而go on doing sth.意为“继续做下去”。如： After reading the text, the students went on to do the exercises.学生们读完课文后，接着做练习。 It’s raining hard, but the farmers go on working on the farm.虽然天正下着大雨，但农民们继续在农场干活。 5. try to do/try doing sth 解析：try to do sth.意为“尽力去做某事”，而try doing sth.意为“（用某一种办法）试着去做某事”。如： Try to come a little early next time, please.下次请尽量早点来。 You can try working out the problem in another way.你可以试试用其它的方法解答这道题目。 6. can’t help to do/can’t help doing sth 解析：can’t help to do为动词不定式结构；can’t help doing sth.意为“身不由己地去做某事”或“情不自禁地去做某事。”如： We can’t help to finish it.我们不能帮忙完成此事。 I couldn’t help laughing when I saw her strange face.当我看到她奇怪的脸时，我情不自禁地笑了。 7. hear sb. do/hear sb. doing sth 解析：hear sb. do sth.意为“听见某人做某事”，指听到了这个动作的全过程；hear sb. doing sth.意为“听到某人做某事”，指听到时候，这个动作正在发生。如： I often hear him sing in the classroom.我经常听见他在教室里唱歌。 Do you hear someone knocking at the door?你听见有人在敲门吗？ 应该说明的是：和hear的用法一样的还有see、watch、notice等。
全部的英语单词的用法与区别 both,either,neither,all,any,none的用法与区别，相信很多同学都傻傻分不清楚。下面是小编为您收集整理的全部的英语单词的用法与区别，供大家参考! 全部的英语单词的用法与区别 这些词都可用作代词或形容词。其位置都在be 动词之后，行为动词之前或第一助动词之后。 1) both (两者都)，either(两者中任何一个)，neither (两者都不)。以上词使用范围为两个人或物。 Neither of the two boys is clever. 两个男孩都不聪明。 2) both，either both与复数连用，either与单数连用。 Both the boys are clever. 两个男孩都很聪明。 Either of the two boys is clever. 两个男孩都很聪明。 There are flowers on both sides of the street. (两岸) There are flowers on either side of the street. (岸的两边) 路边长满了野花。 3) all (所有的，全部的人或物)，any (任何一个)，none (都不)。
以上词使用范围为三者以上。 All the flowers are gone. 所有的花都谢了。 I dont like any of the flowers. 这些花我都不喜欢。 I like none of the flowers. 这些花我都不喜欢。 注意：all与none用法一样。跟单数名词，用单数动词;跟复数名词，用复数动词。 All of the students are there. 所有的学生都在那。 All (of) the milk is there. 所有的牛奶都在那。 in,put on,wear,dress,have on的用法与区别 in，put on，wear，dress, have on, pull on 用法辨析 都含有穿、戴之意，但用法不同。 in是介词，后接表示衣服或颜色的词，着重于服装的款式或颜色。它所构成的短语只能作表语或定语。例如： This is a picture of a young man in a black coat. 这是一张穿着黑色外套的年轻人的照片。这里in a black coat 是young man的定语。 He is in a black nylon jacket today. 今天他穿着黑色尼龙夹克。 In a black nylon jacket在这个句子里作表语。 put on 穿上、戴上，强调穿戴的动作，后接衣服、鞋帽等。例如：
To do 和 doing的用法 1. finish, enjoy, feel like, consider, imagine, keep, postpone, delay, mind, practise, suggest, risk, quit+doing 2. 1）forget to do 忘记要去做某事（此事未做） forget doing忘记做过某事（此事已做过或已发生） 2）stop to do 停止、中断（某件事），目的是去做另一件事 stop doing 停止正在或经常做的事 3）remember to do 记住去做某事（未做） remember doing记得做过某事(已做) 4) regret to do对要做的事遗憾 regret doing对做过的事遗憾、后悔 5）try to do努力、企图做某事 try doing试验、试一试某种办法 6) mean to do打算，有意要… mean doing意味着 7）go on to do 继而（去做另外一件事情） go on doing 继续（原先没有做完的事情） 8）propose to do 打算（要做某事） proposing doing建议（做某事） 9) like /love/hate/ prefer +to do 表示具体行为；+doing sth 表示抽象、倾向概念 （注)如果这些动词前有should一词，其后宾语只跟不定式，不能跟动名词。例如： I should like to see him tomorrow． 10) need, want, deserve +动名词表被动意义；+不定式被动态表示“要(修、清理等)”意思。 Don’t you remember seeing the man before你不记得以前见过那个人吗 You must remember to leave tomorrow．你可要记着是明天动身。 I don’t regret telling her what I thought．我不后悔给她讲过我的想法。(已讲过) I regret to have to do this， but I have no choice．我很遗憾必须这样去做，我实在没办法。(未做但要做) You must try to be more careful．你可要多加小心。 Let’s try doing the work some other way．让我们试一试用另外一种办法来做这工作。 I didn’t mean to hurt your feeling．我没想要伤害你的感情。 This illness will mean (your) going to hospital．得了这种病(你)就要进医院。 3.省to 的动词不定式 1）情态动词 ( 除ought 外，ought to)： 2）使役动词 let, have, make： 3）感官动词 see, watch, look at, notice , observe, hear, listen to, smell, feel, find 等后作宾补，省略to。 注意：在被动语态中则to 不能省掉。 I saw him dance.
重要短语的用法及区别 ★on，in和with. on：表示使用通讯工具、信息或传媒，乘坐交通工具等；I don’t want to talk about it on the phone. in：使用语言文字等媒介；Can you speak it in English? with：借助具体的手段或工具。Don’t write it with a red pen.★at , on , in三者都可以表示“在……的时候”。 At：表示在哪个时刻用；I get up at six o’clock in th e morning . 我早晨六点起床。 On：表示在哪一天，哪一天的早上（下午、晚上）；on Wednesday , on Sunday morning , on May I , on a cold morning in 1936 in：表示在哪一年（季、月），在上午，下午等。in September , in the morning , in the afternoon ★spend,pay,cost,take
Sb. spend …on sth. 某人花了…（时间、金钱）在某事上。 (in) doing sth. 某人花了…（时间、金钱）做某事。Sb. pay …for sth.某人为某物花了…钱。 Sth. cost sb. …某物花了某人…钱。 It takes/took sb. … to do sth.花了某人…（时间、金钱）做某事。 ★too much, too many, much too too much + 不可数名词too many + 可数名词much too + 形容词 ★not … until &u ntil not … until直到…才… （主句动词是短暂性动词） until 一直到…（主句中使用延续性动词） ★few, a few; little , a little. 虽然都表示“少”，但 （1）few, a few是可数的, little, a little是不可数的。 （2）a few, a little含肯定意味，few, little含否定意味。
a用在以辅音字母开头，或以读做辅音的元音字母开头的单词前面： a man一个男人 a university一所大学 a hat一顶帽子 a European一个欧洲人 a one-way street一条单行马路 an用在以元音字母（a，e，i，o，u）开头，或以不发音的h字母开头的单词前面：an apple一个苹果 an island一个岛 an uncle一位大叔 an onion一个洋葱 an egg一个鸡蛋 an hour一小时 an还用在发音以元音开头的单个字母前面： an L－plate一块“实习驾驶”车牌 an MP一个国会议员 an SOS一个呼救信号 an…x?一个x字母、X形的东西或未知数 a／an没有性的变化： a man一个男人 a woman一个女人 an actor一个男演员 an actress一个女演员
a table一张桌子 a／an不用在以下几种情况下： A 复数名词之前： a／an没有复数形式，所以a dog的复数是dogs，an egg的复数是eggs。 B 不可数名词之前（参见第13节）： C 三餐名称之前，但这些名称前加形容词时除外： We have breakfast at eight． 我们8点钟吃早饭。 He gave us a good breakfast． 他请我们吃了一顿丰盛的早餐。 为了庆祝某事或特意为某人而举行的宴会之前要用冠词： I was invited to dinner（at their house， in the ordinary way）． 他们邀请我吃饭（在他们家吃便饭）。 但是说： I was invited to a dinner given to welcome the new ambassador． 我被邀请参加欢迎新任大使的宴会。 A a／an和one（形容词） 1 计算时间、测量距离或重量等时，a／an或one可以用于单数的前面：￡l=a／one pound一英镑 ￡ 1，000，000＝ a／one million pounds 100万英镑 （参见第三十六章。） 但是，请注意： The rent is ￡ 100 a week． 房租为每星期100英镑。
一．To do形式 afford to do sth. 负担得起做某事 agree to do sth. 同意做某事 arrange to do sth.安排做某事 ask to do sth. 要求做某事 beg to do sth. 请求做某事 care to do sth. 想要做某事 choose to do sth. 决定做某事 decide to do sth. 决定做某事 demand to do sth. 要求做某事 expect to do sth. 期待做某事 fear to do sth. 害怕做某事 help to do sth. 帮助做某事 hope to do sth. 希望做某事 learn to do sth. 学习做某事 manage to do sth. 设法做某事 offer to do sth. 主动提出做某事 plan to do sth. 计划做某事 prepare to do sth. 准备做某事 pretend to do sth. 假装做某事 promise to do sth. 答应做某事 refuse to do sth. 拒绝做某事 want to do sth. 想要做某事 wish to do sth. 希望做某事 happen to do sth. 碰巧做某事 struggle to do sth. 努力做某事 advise sb. to do sth. 建议某人做某事allow sb. to do sth. 允许某人做某事 ask sb. to do sth.请（叫）某人做某事bear sb. to do sth.忍受某人做某事 beg sb. to do sth. 请求某人做某事cause sb. to do sth. 导致某人做某事command sb. to do sth. 命令某人做某事drive sb. to do sth .驱使某人做某事elect sb. to do sth. 选举某人做某事encourage sb. to do sth. 鼓励某人做某事expect sb. to do sth. 期望某人做某事forbid sb. to do sth. 禁止某人做某事force sb. to do sth. 强迫某人做某事 get sb. to do sth. 使（要）某人做某事hate sb. to do sth. 讨厌某人做某事 help sb. to do sth. 帮助某人做某事
常用英语词语用法辨析 1:any ，some二者都可作"一些"讲。any只用于否定句和疑问句，some一般用于肯定句，但可用于表示请求，建议、反问等的疑问句中。例如：Have you any ink ? 你有墨水吗? We have to learn some English . 我们得学点英语。 Will you lend me some English books ? 你肯借给我几本英语书吗? 2:around ，round二者作为介词和副词都有"在周围"的意。其区别在于around多用美国，round多用于英语。例如：The earth goes round (或around)the sun . 地球绕太阳运行。 3:arrive ，reach ，get to三者都有"到达"之意。reach为及物动词，arrive为不及物动词，后面接介词in或at.get to常用于口语，可代替前二者。例如：The foreign visitors have arrived in Beijing . 外宾已到达北京。 She arrived at her hometown yesterday . 昨天她到达了她的家乡。 They reached Tianjin yesterday . 昨天他们到达天津。 4:as ，since ，because ，for作为连词，它们都有"因为"的意思。但because语气最强，as和since其次，for是并列连词，常用来说明理由或提供一种解释。例如：Because I am too busy ，I can't do it now . 因为我太忙，所以我现在不能做此事。 As he did not know the way ，he went to ask an old man .因为不认识路，他去问一位老人。 Since his method doesn't work ，let's try yours . 既然他的方法不行，我们就试试你的吧。 It will rain ，for it's getting dark . 要下雨了，因为天色都暗了。 5:as ，when ，while作为连词，都有"当……的时候"，其区别在于用when时，从句的动作和主句的动作可以同时，也同先于主句的动作，用as时，主句和从句的动作同时发生，有时可译作"一边……，一边"，用while常表示一段较长的时间或一个过程，主句和从句的动作同时发生。例如：When the teacher came in ，the students stood up . 当老师进来时，学生们起立。 As the children walked along the lake ，they sang merrily .他们一边沿着湖走，一边愉快地唱歌。 Don't talk while you are eating . 吃饭时你不要说话。 6:be asleep ，go to sleep二者都表示"睡着"，但前者表示"睡着"的状态，后者表示"入睡"的过程。例如：While I was watching TV ，he was asleep . 我在看电视的时候，他睡着了。 I didn't know when I went to sleep . 我不知道我什么时候入睡的。 7: at ，on ，in三者都可以表示"在……的时候"区别在于：表示在哪个时刻用at;表示在哪一天，哪一天的早上(下午、晚上)，用on;表示在哪一年(季、月)。在上午，下午等，用in.例如：I get up at six o'clock in the morning . 我早晨六点起床。 on Wednesday ，on Sunday morning ，on May I ，on a cold morning in 1936 ，in autumn ，in September ，in the morning ，in the afternoon 22. battle ，fight ，struggle它们作为名词，都有"战斗"的意思。battle和fight在"战斗"这个意义上，有时可以通用。例如：The battle(fight)lasted a long time . 这次战斗持续了很长时间。还常指战争中的一次战役。例如：the Battle of Waterloo 滑铁卢战役fight的意义比较广泛。例如：We have started a fight against pollution . 我们已开始了一场清除污染的斗争。 struggle有"斗争"，"奋斗"的意思。例如：class struggle阶级斗争，struggle for independence 为独立而斗争，struggle with a person跟某人作斗争。 8:beautiful ，pretty ，handsome三者都有"美"的意思。但含义及所形容的对象不同。
识记:初中英语非谓语动词总结（中考常考） 记住：动词后面加动名词表示已经做了；加动词不定式表示将要去做。 记住：动词后面加动名词表示经常做；加动词不定式表示一次做。 * *跟动词原形的词有：“一感二听三让四看”，即：feel, // hear, listen to, // let, make, have,// look at, see, wact notice.// 一．后面可跟动词的ing形式的情况。 1.动词：*以下记住每一个词组的第一个动词。 finish doing sth.完成做某事；enjoy doing sth. 喜欢做某事； practice doing sth. 练习做某事；imagine doing,想象做某事； avoid doing sth.避免做某事；consider doing sth.考虑做某事； suggest doing sth.建议做某事；mind doing sth.介意做某事； * keep doing sth.持续做某事, miss doing错过做, advise doing建议做；* keep sb doing让某人一直做 2.固定短语： feel like doing sth.喜欢做某事；be busy doing sth.忙于做某事； be worth doing 值得做某事；spend time (in) doing sth.花费时间（金钱）做某事； have difficult/trouble in doing sth做某事有困难；have fun doing.做某事高兴 3. 介词后(on, in, of, about, at, with, without, for, from, up, by等)： 如：be good at doing sth.；thank you for doing sth.；give up doing sth.；stop sb. from doing sth.；do well in do sth.； be afraid of doing sth.；be interested in doing sth.；be proud of；instead of；be fond of；what/how about doing sth?某事怎么样？ 4. to作介词时，后跟动名词的情况： look forward to doing sth期望做某事；prefer doing sth. to doing sth与…相比较更喜欢…； pay attention to doing注意做某事；be/get used to doing sth.习惯于做某事； make a contribution to doing为…做贡献 No+动名词，表示禁令No smoking禁止吸烟No parking禁止停车 5. go+动名词，意思是去进行某种活动或运动： go shopping,去购物；go skating,去滑冰；go hiking去远足（旅行） 6. do some/the+动名词，指进行某种活动： do some cleaning,搞卫生；do some washing 洗衣服； 二．后面可跟动词的不定式形式的情况。 1.动词：不需要记住哪些动词后跟动词不定式。 2.句型：(1)动词： allow sb. to do sth. 允许某人去做某事（区分allow doing sth） ask sb. (not) to do sth. 叫某人做事某事（叫某人不要去做某事） tell sb. (not) to do sth. 叫某人去（不要）做某事follow sb. to do sth. 跟随某人去做某事 get sb. to do sth. 让某人去做某事warn sb. (not) to do sth. 警告某人做某事（或不要做某事） encourage sb to do鼓励某人做、expect sb to do期待某人做 invite sb to do邀请某人做、teach sb to do教会某人做 advise sb to do建议某人做（区分advise / suggest doing sth） (2) Be+形容词adj.（即：情感类的形容词）+ to do
英语中常见可分可合单词的用法区别 在英语中，有时两个部分合在一起构成一个词，表达的是一种意思；两个部分分开，各成一个词时，表达另一种意思，有另外的用法。现将中学英语课本中出现的这类“分合有别、用法各异”的词语归纳分析如下： 1. away 与 a way away 用作副词时，意为“远”、“离开”、“远离”，通常作be 的表语，也可以放在动词或副词之后，与动词一起构成短语动词；a way 是名词短语，是“一条路”、“一个方法”之意。例如： Who will look after these children when I'm away? 我走了之后，谁来照看这些孩子？ The village is about eight miles away from here. 那个村子离这儿大约有八英里远。 She thought hard，and she found a way at last. 她苦思冥想，终于找到了一个方法。 2. already 与all ready already 是副词，意为“已经”，常与完成时态连用；all ready 是形容词短语，常用作表语，意为“一切就绪”、“都准备好了”。例如： We have already finished our work. 我们已经做完了我们的工作。 We are all ready for the trip. Let's go. 我们都做好了旅行的准备。那么咱们就上路吧。 3. awhile 与a while awhile 是副词，意思是“一会儿”、“片刻”，其前不可加介词或副词； a while 是名词短语，意思是“一会儿”(与awhile 同义)，a while 前面可以加介词in，after或for，也可加副词quite。例如： Wait awhile. [= Wait (for) a while. ] 等一会儿。 The postman came quite a while ago. 邮递员好一会儿前来过。 After a while, they quarrelled again. 过了一会儿，他们又吵了起来。 4. asleep 与a sleep