Java versus .Net - PowerPoint PPT Presentation

java vs net l.
Skip this Video
Loading SlideShow in 5 Seconds..
Java versus .Net PowerPoint Presentation
Java versus .Net

play fullscreen
1 / 54
Download Presentation
Download Presentation

Java versus .Net

Presentation Transcript

  1. Java vs. .Net Bent Thomsen Department of Computer Science Aalborg University

  2. A typical.NET Enterprise Solution IIS on W2k Server SQL Server Browser .NET managed component ASP .NET Windows Client

  3. A typical J2EE Enterprise Solution Java App Server DB Server Browser Servlet JSP EJB Java Client

  4. Java vs .Net Solutions • Both multi-tiered, similar computing technologies • Both support “standards” • Both offer different tools & ways to achieve the same goal. • A lot of parallelism can be seen. • Very difficult to compare and qualify the comparison because each has its own advantages & disadvantages.

  5. The TMC Petshop Performance Case Study • Java Pet Store is Sun’s primary blueprint application for J2EE • Source: • Illustrates best coding practices for J2EE • Ships as a sample application in IBM Websphere, Oracle Application Server 9i, Sun iPlanet, and BEA WebLogic • The .NET Petshop is a port of the J2EE Java Pet Store to .NET • Source: • Implements the same functionality as Java Pet Store • Illustrates best coding practices for .NET Framework • In the TMC Petshop Performance Case Study, The Middleware Company implemented both the Java Pet Store and the .Net Petshop. • The J2EE version ran on two different application servers • All versions used the same hardware and OS

  6. Java Pet Store Components • The Storefront presents the main user interface in a Web front-end. Customers use the Storefront to place orders for pets. • The Order Processing Center (OPC) receives orders from the Storefront. • The Supplier fulfills orders from the OPC from inventory and invoices the OPC. • The Admin presents the administrator interface in a JFC/Swing front-end. Administrators use the Admin to examine pending orders and approve or deny them.

  7. Java Pet Store vs. .Net Pet Shop

  8. Porting Java Pet Store to .NET 15500 Lines of Code Required 14,273 14000 .NET Petshop 11500 Java Pet Store 9000 7500 5,891 5,404 4,410 5000 2,865 2,566 2500 710 761 412 74 Total Lines of Code User Interface Middle Tier Data Tier Configuration

  9. TMC Pages per Second

  10. TMC Max Supported Users

  11. How Microsoft interprets the data • Based on Oracle-published data for tuned version of Java Pet Store • Using Oracle’s test scripts from their “9i App Server Challenge” • Run on equivalent hardware Supporting 6 times more users 1.0 0.8 0.6 2800% Better performance Response Time (Seconds) 0.4 0.2 0 0 500 1000 1500 2000 2500 2750 User Load Level

  12. The “Conclusions” • Microsoft concludes: ".NET is 28 times faster than J2EE" • Because: • “Java PetShop” is Sun’s “best practice” example • Numbers compared with published Oracle results • Identical functionality, all code published and documented • Can we believe the raw numbers? Yes! • Why? Microsoft delivers full docs on the entire scenario • But 28 times really?

  13. The Shootout afterwards • Java PetShop: J2EE blueprint application • Built by Sun to show "best practice" design • Implementation by TMC missed quite a few optimizations • .NET PetShop • Built by TMC (with help from MS) with different design • Plain classes instead of container managed components • Middle tier moved into ASP.NET • Using stored procedures instead of ad-hoc SQL • Uses server-side caching of ASP.NET • Many performance optimizations applied • Most Java devotees find this highly unfair

  14. TMC revisits the Pet Shop • Re-implementation of J2EE version • 17 times performance increase • Second version showed some J2EE implementation equal .Net • Second version is a testimony to performance tuning

  15. What does the comparison tell us? • It is very difficult to make such comparisons • That .Net has gained maturity extremely fast • That the two frameworks are very similar • The Devil is in the detail So let’s look at some details

  16. Comparing the stacks VB C++ C# Perl Python … … Eclipse Webshpere Studio Win32 MSMQ, COM+, IIS, WMI, AD, ADAM, Indexing, UDDI, etc. BEA Weblogic Enterprise solutions Third party extensions Extensions: Phoenix, Tiles, Java Faces P&P blocks Struts ASP.Net JSP Servlets Visual Java JDBC ADO.NET J2EE Class Library Base Class Library Java runtime CLR J2EE App Servers Websphere, Weblogic, Tomcat, etc. JMS Apache Win32, Unix, Linux

  17. Java vs. C# • C# is an object oriented language of the C++/Java flavor • Syntax similar to Java and C/C++. • Quite an impressive design and care for details • Java developers will feel comfortable • most of the time and frustrated when things are different • MS says: “C# combines the power of VC++ with the ease of usage of VB” • Not really true: • C# is really powerful BUT • It is not easy to learn for non C++/Java programmers • It is the language to learn if you are serious about .NET!

  18. Java vs. C# // This is a comment in Java code class HelloWorld{ public static void main(String[] args){ for(int i= 1; i<= 100; i++) System.out.println("Hello!");  } } // This is a comment in C# using System; class HelloWorld{static void Main(){ for(int i=1; i<=100; i++)     Console.WriteLine("Hello");  }} }

  19. Primitives Java int, float, double, etc. Allocated on stack Not an Object Not extensible C# int, float, double, etc. structs Allocated on stack Inherited from object class structs can implement interfaces Cannot inherit from another class // C# struct Point { int x; int y; Point(int x, int y) { this.x = x; this.y = y; } }

  20. Classes – Overriding Methods Java All methods are implicitly virtual C# Explicitly use ‘virtual’ and ‘override’ keywords class B { public void foo() { } } class D extends B { public void foo() { } } // D’s foo() overrides B’s foo() class B { public virtual void foo() { } } class D:B { public override void foo() { } } // D’s foo() overrides B’s foo()

  21. Exceptions Java C++-style try/catch blocks finally– action done even after an exception is caught throws – methods identify what exceptions they throw C# C++-style try/catch blocks finally – same as Java Does not support throws clause // Java and C# try { // Stuff… } catch { // Ack! } finally { // Always! } // Java – throws an IOException public void myFunc(int a) throws IOException { // Work… }

  22. Java vs. C#

  23. Java vs. C# vs. C++ vs. VB

  24. Java vs. C# in High-performance computations

  25. C# Features not in Java • No automatic fall-through from one case block to the next • Strongly-typed enums • By reference calls are explicit at caller AND callee • Method overrides are explicit • Supports versioning • Structs (value types) • Integrated support for properties and events • Can still use pointers with RAD language • Can share data and use functionality with components written in many different languages

  26. CLR vs JVM C# VB .Net Managed C/C++ Lots of other Languages Java MSIL Byte Codes CLR CTS GC Security Runtime Services JRE (JVM) GC Security Runtime Services Windows OS Mac Win Unix Linux Both are ‘middle layers’ between an intermediate language & the underlying OS

  27. .Net on other platforms • ECMA standardisation • ECMA 334 and 335 • The Mono Project • Shared Source Common Language Runtime • CLI, C# and Jscript • FreeBSD and Windows implementations • Linux port underway • DOT GNU project • Portable .Net implementation • C# for both CIL and JVM

  28. JVM vs. CLR • JVM designed for platform independence • Single language: Java (?) • A separate JVM for each OS & device • CLR designed for language independence • Multiple languages for development • C++, VB, C#, (J#) • APL, COBOL, Eiffel, Forth, Fortran, Haskel, SML, Mercury, Mondrian, Oberon, Pascal, Perl, Python, RPG, Scheme, SmallScript, … • Impressive usage of formal methods and programming language research during development • Impressive extensions for generics and support for functional languages underway • Underlying OS: Windows (?)

  29. Java Byte Code and MSIL • Java byte code (or JVML) is the low-level language of the JVM. • MSIL (or CIL or IL) is the low-level language of the .NET Common Language Runtime (CLR). • Superficially, the two languages look very similar. • One difference is that MSIL is designed only for JIT compilation. • The generic add instruction would require an interpreter to track the data type of the top of stack element, which would be prohibitively expensive. MSIL: ldloc.1 ldloc.2 add stloc.3 JVML: iload 1 iload 2 iadd istore 3

  30. JVM vs. CLR at a glance

  31. J2EE 1.5 • J2EE (1.5) preview of 26.4.2004 • Focus on ease of development • Generics and metadata as in J2SE 1.5 (more like C#) • Java Studio Creator tool (in beta from April 2004) (more like Visual Studio .Net) • Timeframe • To be discussed at JavaOne in June • Finalized in approximately one year • IBM push for truly open source Java • Others hesitate (even Open Source JBoss)

  32. What about mobile? • Web Services are fine as n-tier applications with UI provided through browser, but … • On mobile devices WAP hasn’t really caught on • Even iMode hasn’t caught on in Europe • Renewed Thin/Thick client discussion • Java applications on Mobile devices are reasonably successful • Now Microsoft is moving fast into the field with .Net Compact Framework

  33. Two Devices Nokia 6600 Orange SPV E200

  34. Two Devices

  35. Building Mobile Solutions is harder • Mobile device • Software infrastructure, hardware requirements • Communication technology • On-/Offline scenario • Wireless Wide Area Networks/ Wireless Local Area Networks • Communication protocol • Application architecture scenario • Thin/fat client • Data management • Synchronisation • On-/offline capabilities • Security issues • Dangers for mobile devices • Threats of communication technology • Danger of exposing enterprise data

  36. Therefore ... The Java vs. Net discussion goes mobile • Java 2 Micro Edition (J2ME) is not ONE Java edition • An J2ME compliant application consists of • Configuration • Profile (e.g. Personal, Mobile Information Device Profile (MIDP)) • Application code • Vendor specific UI • 3 Contenders to compare • Java 2 Micro Edition – Connected Device Configuration (CDC) • Java 2 Micro Edition – Connected Limited Device Configuration (CLDC) • Microsoft .NET Compact Framework

  37. Student Project • Group of Master Level Students (Hovedfag) • Bjørn D. Rasmussen • Casper S. Jensen • Jimmy Nielsen • Lasse Jensen • Collaboration with DEIF • Project Goals • Build end-to-end mobile client, wireless, webservices based application with back-end interface to SCADA • In Java (J2ME/J2EE) and in .Net • Compare the two solutions on a set of criteria

  38. Basis of comparison Objective measurements Lines of code Development tools Documentation Infrastructure Performance Architectural pattern Security Price Development time Subjective measurements Developer satisfaction End-user satisfaction

  39. DEIF M-Vision (SCADA up and running in 30 minutes)

  40. SCADA on SmartPhones

  41. Development tools and Documentation • Server-side is well supported by both Java and .NET IDEs • On the client-side .NET IDEs benefit from the fact that .NET CF is so close to .NET (With Java there are separate IDEs for desktop and mobile application development) • Compatibility problems between Java vendors • Java IDEs are slow! • C# is a richer/more complex language than Java • Both Java and .NET have well documented API • Web service documentation • .NET - MSDN • Java – Google • Support for encryption of web services • .Net CF: HTTPS and SOAP extensions • J2ME: HTTPS, but only in CDC & MIDP 2.0

  42. Performance • Server-side web service performance study with: • Microsoft Application Server (v. 5.1) • Sun ONE Application Server (v. 7.0) • Tested for: • Throughput • Failure rate • Average response time

  43. Performance - Throughput

  44. Performance – Average response time

  45. Performance – Failure rate

  46. Development time • Slow start-up when developing in Java • Jungle of web service implementations • IDE incompatibility problems • Emulators • kSOAP • Trouble-free implementation in .NET

  47. Subjective Measures • Developer satisfaction • VS.NET integrates web service and mobile application development far better than any Java IDE • A subset of an API is a better solution for mobile devices instead of an entirely new API • End-user satisfaction • DEIF would choose a .NET based solution since startup time is very important • DEIF only needs a limited number of IDE licenses • Extra price for a SmartPhone 2003 is insignificant compared to the value it gives • The SmartPhone 2003 is more ”fancy” than the Java phones

  48. Students Conclusions • We see .NET as a better designed framework because: • it eliminates the language barrier while also being platform independent • it makes only little distinction between desktop and mobile application development • Sun’s application server performance is very poor compared to IIS. • License fees for a Java based solution are cheaper but .NET might catch up when considering development time • We tried a combined .NET and Java solution but this has shown to be very troublesome!

  49. My conclusions • The two worlds are becoming more and more similar • But it seems that you have to work harder in Java than in .Net • .Net is currently driving technology transfer from Research to Product • Windows generation at University now • Watch-out in the mobile world • Vodafone to offer Microsoft Smart phones • • Fed-up with Nokia promoting own brand, rather than operator brand

  50. What do these comparisons tell us? • It is very difficult to make such comparisons • That .Net has gained maturity extremely fast • That the two frameworks are very similar • You will not be sacked for choosing the right J2EE application server ;-) • The Devil is in the detail • C# is not Java • ADO.NET is not JDBC • CLR is not JVM • CF.Net smartphones are very different from Java Smartphones