From long time one reason for me to update to newer Java version was always bug fix and performance improvement. Apart from major changes like Generics in Java 1.5 and Lambdas in Java 8, there are so many small improvements, performance optimization which just goes under radar, one of such change is creating empty ArrayList and HashMap with size zero in JDK 1.7.0_40 update. Many Java developer doesn't even know about these changes, part of the blame lies on Java developers like me, as I hardly read release notes of minor Java updates. Some times these changes are done as part of bug fixes and other time as minor optimization, but given popularity of ArrayList and HashMap in Java application impact of this simple Java optimization is huge. If you are running on Java 1.6 or earlier version of Java 1.7, you can open code of java.util.ArrayList and check that, currently empty ArrayList is initialized with Object array of size 10. If you create several temporary list in your program, which remains uninitialized, due to any reason then you are not only losing memory but also losing performance by giving your garbage collector more work. Same is true for empty HashMap, which was initialized by default initial capacity of 16. This changes are result of observation made by Nathan Reynolds, and Architect at Oracle, which apparently analysed 670 Java heap dumps from different Java programs to find out memory hogs.
Wednesday, July 23, 2014
Thursday, July 17, 2014
TCP and UDP are two transport layer protocols, which are extensively used in internet for transmitting data between one host to another. Good knowledge of how TCP and UDP works is essential for any programmer. That's why difference between TCP and UDP is one of the most popular programming interview question. I have seen this question many times on various Java interviews , especially for server side Java developer positions. Since FIX (Financial Information Exchange) protocol is also a TCP based protocol, several investment banks, hedge funds, and exchange solution provider looks for Java developer with good knowledge of TCP and UDP. Writing fix engines and server side components for high speed electronic trading platforms needs capable developers with solid understanding of fundamentals including data structure, algorithms and networking. By the way, use of TCP and UDP is not limited to one area, its at the heart of internet. The protocol which is core of internet, HTTP is based on TCP. One more reason, why Java developer should understand these two protocol in detail is that Java is extensively used to write multi-threaded, concurrent and scalable servers. Java also provides rich Socket programming API for both TCP and UDP based communication. In this article, we will learn key differences between TCP and UDP protocol, which is useful to every Java programmers. To start with, TCP stands for Transmission Control Protocol and UDP stands for User Datagram Protocol, and both are used extensively to build Internet applications.
Monday, July 14, 2014
Thursday, July 10, 2014
An Armstrong number of three digits is an integer such that the sum of the cubes of its digits is equal to the number itself. For example, 153 is an Armstrong number, since 1**3 + 5**3 + 3**3 = 153, 371 is an Armstrong number since 3**3 + 7**3 + 1**3 = 371. Along with usual beginner exercises e.g. calculating factorial, reversing string or calculating prime numbers, this is a good exercise to build programming logic. It teaches you basic programming technique of how to use operator for something which is not obvious, for example, to solve this programming challenge, we first need to check if a number is Armstrong or not, and to do this we need individual digits of the number. how do we do that? well there is a programming technique, which you might have learned while doing number palindrome exercise. If you modulo an integer number by 10, you will get last digit, for example 656%10 will give you 6, which is last digit of 656. Similarly to reduce the input after each iteration, you can use division operator, as 656/10 will return 65, which is without last digit. If you know this trick, it's very easy to solve any programming problem, which requires examination of individual digits. This Java program uses same technique and compute all Armstrong numbers in the range of 0 and 999. By the way this program has different variations as well e.g. how do you find Armstrong number of four digit, as this program only calculates three digit Armstrong number. Well, for that you need to remember general definition of Armstrong number which is, An Armstrong number is an n-digit number that is equal to the sum of the nth powers of its digits. So a four digit Armstrong number will be equal to sum of power four of individual digits of that number. Another interesting challenge is to write a program which uses recursion to find if number is Armstrong or not.
Tuesday, July 8, 2014
Why Java developer should read a book on Performance tuning? When I first faced this question long time back, I thought I will do it later, but I never get back to that for a long time. I realize my mistake of having lack of knowledge on performance measurement, tuning and finding bottleneck only when I faced serious performance and scalability issues on our mission critical server side financial application written in Java. It's true that when you really need it you learn most, but those times are not the best time to learn fundamentals, in fact those times are to apply and correct your misunderstanding. This is why I am sharing these Java performance books to all Java programmers and suggesting them to take sometime and go through at-least one book in full. By the way these books are in addition to my 5 must read books for Java programmers. Remember knowledge of Performance tuning is one important aspect of Senior Java developers, and can separate you from crowd. Ever since Java was introduced, almost 20 years back, it has faced criticism on being slow and lacking performance. Today, I don't think Java is anywhere behind in terms of performance to native languages. Given Java's ability to natively compile hot code using JIT (Just in time Compiler), it is almost at par with native applications written in C and C++, but a lot can be done by following best practices, avoiding common performance pitfalls and using latest tools and techniques. In this article, I am going to introduce with five + one good books on Java performance, which will not only teach you what to measure, how to measure but also explains fundamentals and concept behind those issues. You will not only learn about System and JVM on which your Java application run but also how to to writer faster coding using Java API. So what are we waiting for, let's begin our journey to land of great books on Java performance tuning.
Wednesday, July 2, 2014
You go to any Java interview, senior or junior, experience or freshers, you are bound to see couple of questions from thread, concurrency and multi-threading. In fact this built-in concurrency support is one of the strongest point of Java programming language and helped it to gain popularity among enterprise world and programmers equally. Most of lucrative Java developer position demands excellent core Java multi-threading skills and experience on developing, debugging and tuning high performance low latency concurrent Java applications. This is the reason, it is one of the most sought after skill on interviews. In a typical Java interview, Interviewer slowly starts from basic concepts of Thread by asking questions like, why you need threads, how to create threads, which one is better way to create threads e.g. by extending thread class or implementing Runnable and then slowly goes into Concurrency issues, challenges faced during development of concurrent Java applications, Java memory model, higher order concurrency utilities introduced in JDK 1.5, principles and design patterns of concurrent Java applications, classical multi-threading problems e.g. producer consumer, dining philosopher, reader writer or simply bounded buffer problems. Since its not enough just to know basics of threading, you must know how to deal with concurrency problems e.g. deadlock, race conditions, memory inconsistency and various thread safety related issues. These skills are thoroughly get tested by presenting various multi-threading and concurrency problems. Many Java developers are used to only look and read interview questions before going for interview, which is not bad but you should not be too far away. Also collecting questions and going through same exercise is too much time consuming, that's why I have created this list of top 50 Java multi-threading and concurrency related questions, collected from various interviews. I am only going to add new and recent interview questions as and when I am going to discover them. By the way, I have not provided answers of this questions here, Why? because I expect most of Java developer to know the answers of this question and if not, also answers are widely available by using Google. If you don't find answer of any particular question, you can always ask me in comments section. You can even find answers of few questions on the link provided or my earlier post Top 12 Java Thread Questions with Answers.