How to query on blockchain

7. The reason is that the three-phase load should be balanced as far as possible. That is to say, the total power of single-phase electrical appliances connected to the three live wires should be balanced as far as possible to avoid the situation that the power of electrical appliances connected to one phase is too much or too little. In addition, when there are three-phase motor equipment, Different phase of three wire connection may cause the motor to turn in the opposite direction, so the phase of three-phase socket should be ensured as far as possible, that is, the position of three live wires on the socket is consistent.

8. In the actual development, we often encounter some situations like this,
(1) in the case of sudden high concurrency, when the server is under great pressure and calls some methods for more than 100ms and does not respond, we should automatically reject the service instead of blocking until the server crashes, It can be regarded as a disguised service degradation
(2) for the unpredictable code that may have deadlock, add time threshold limit to avoid unlimited resource competition
(3) for the unpredictable code that may have deadlock, add time threshold monitoring, Avoid dead cycle deterioration

some mature open source frameworks generally have the function of timeout response, such as
search framework: Lucene, Solr, elasticsearch,
RPC framework: Dubbo, thrift,
service coordination framework: zookeeper, etc.
if we want to add this function, what should we do
in fact, it's not difficult to add a timing function before and after the method call. If the threshold is exceeded, the thread should be terminated actively and the exception code should be thrown. Of course, in order not to affect the operation of the main service, this code should run independently in a sub thread, rather than coupled in the main thread. The idea is roughly the same. Of course, in order not to build a wheel repeatedly, we suggest to use Google's open source guava toolkit, which provides many useful functions, including the automatic termination of timeout method introced this time, and some functions to limit network traffic or the number of visits within a specified period of time, You can see guava's documentation for yourself<
timelimiter is an interface. There are two subclasses below,
faketimelimiter, which is often used to limit time-out debugging ring debugging.
simpletimelimiter is often used in formal methods, which call method time-out, that is, throw exception.
this paper focuses on the use of simpletimelimiter, There are mainly two methods in this class:
(1) newproxy implements timeout interception through JDK dynamic proxy and callwithtimeout
(2) callwithtimeout implements timeout interception through callable callback
the difference between the two methods lies in that the strategy implemented through proxy can be applied to every called method in this class, The implementation of timeout interception
and the implementation strategy through callback is applicable to the implementation of timeout interception only for a certain code block or method
introce the latest guava package through Maven:
java code

& lt; dependency> < br /> < groupId> com.google.guava</ groupId> < br /> < artifactId> guava</ artifactId> < br /> < version> 19.0</ version> < br /> </ dependency> <
A: Interception Based on proxy mode:
since guava uses JDK dynamic proxy to implement AOP interception, it requires that the proxy class must have an interface implementation, so it's a little troublesome to use.
we need to define an interface first, then define the implementation class, and rewrite the method to be called:
b: callback based usage, which is relatively simple, Because there is no need to use proxy mode,
both methods can implement method call timeout interruption, proxy mode is suitable for class level method timeout interruption, while callback based mode is more suitable for any single method.