Search Results

Search found 8397 results on 336 pages for 'implementation'.

Page 39/336 | < Previous Page | 35 36 37 38 39 40 41 42 43 44 45 46  | Next Page >

• WebBrowserControl for the .NET Framework 1.1

  An implementation of a web browser control for the .NET Framework 1.1

  Read the article

• How to reserve public API to internal usage in .NET?

  - by mark

  Dear ladies and sirs. Let me first present the case, which will explain my question. This is going to be a bit long, so I apologize in advance :-). I have objects and collections, which should support the Merge API (it is my custom API, the signature of which is immaterial for this question). This API must be internal, meaning only my framework should be allowed to invoke it. However, derived types should be able to override the basic implementation. The natural way to implement this pattern as I see it, is this: The Merge API is declared as part of some internal interface, let us say IMergeable. Because the interface is internal, derived types would not be able to implement it directly. Rather they must inherit it from a common base type. So, a common base type is introduced, which would implement the IMergeable interface explicitly, where the interface methods delegate to respective protected virtual methods, providing the default implementation. This way the API is only callable by my framework, but derived types may override the default implementation. The following code snippet demonstrates the concept: internal interface IMergeable { void Merge(object obj); } public class BaseFrameworkObject : IMergeable { protected virtual void Merge(object obj) { // The default implementation. } void IMergeable.Merge(object obj) { Merge(obj); } } public class SomeThirdPartyObject : BaseFrameworkObject { protected override void Merge(object obj) { // A derived type implementation. } } All is fine, provided a single common base type suffices, which is usually true for non collection types. The thing is that collections must be mergeable as well. Collections do not play nicely with the presented concept, because developers do not develop collections from the scratch. There are predefined implementations - observable, filtered, compound, read-only, remove-only, ordered, god-knows-what, ... They may be developed from scratch in-house, but once finished, they serve wide range of products and should never be tailored to some specific product. Which means, that either: they do not implement the IMergeable interface at all, because it is internal to some product the scope of the IMergeable interface is raised to public and the API becomes open and callable by all. Let us refer to these collections as standard collections. Anyway, the first option screws my framework, because now each possible standard collection type has to be paired with the respective framework version, augmenting the standard with the IMergeable interface implementation - this is so bad, I am not even considering it. The second option breaks the framework as well, because the IMergeable interface should be internal for a reason (whatever it is) and now this interface has to open to all. So what to do? My solution is this. make IMergeable public API, but add an extra parameter to the Merge method, I call it a security token. The interface implementation may check that the token references some internal object, which is never exposed to the outside. If this is the case, then the method was called from within the framework, otherwise - some outside API consumer attempted to invoke it and so the implementation can blow up with a SecurityException. Here is the modified code snippet demonstrating this concept: internal static class InternalApi { internal static readonly object Token = new object(); } public interface IMergeable { void Merge(object obj, object token); } public class BaseFrameworkObject : IMergeable { protected virtual void Merge(object obj) { // The default implementation. } public void Merge(object obj, object token) { if (!object.ReferenceEquals(token, InternalApi.Token)) { throw new SecurityException("bla bla bla"); } Merge(obj); } } public class SomeThirdPartyObject : BaseFrameworkObject { protected override void Merge(object obj) { // A derived type implementation. } } Of course, this is less explicit than having an internally scoped interface and the check is moved from the compile time to run time, yet this is the best I could come up with. Now, I have a gut feeling that there is a better way to solve the problem I have presented. I do not know, may be using some standard Code Access Security features? I have only vague understanding of it, but can LinkDemand attribute be somehow related to it? Anyway, I would like to hear other opinions. Thanks.

  Read the article

• Haskell: Left-biased/short-circuiting function

  - by user2967411

  Two classes ago, our professor presented to us a Parser module. Here is the code: module Parser (Parser,parser,runParser,satisfy,char,string,many,many1,(+++)) where import Data.Char import Control.Monad import Control.Monad.State type Parser = StateT String [] runParser :: Parser a -> String -> [(a,String)] runParser = runStateT parser :: (String -> [(a,String)]) -> Parser a parser = StateT satisfy :: (Char -> Bool) -> Parser Char satisfy f = parser $ \s -> case s of [] -> [] a:as -> [(a,as) | f a] char :: Char -> Parser Char char = satisfy . (==) alpha,digit :: Parser Char alpha = satisfy isAlpha digit = satisfy isDigit string :: String -> Parser String string = mapM char infixr 5 +++ (+++) :: Parser a -> Parser a -> Parser a (+++) = mplus many, many1 :: Parser a -> Parser [a] many p = return [] +++ many1 p many1 p = liftM2 (:) p (many p) Today he gave us an assignment to introduce "a left-biased, or short-circuiting version of (+++)", called (<++). His hint was for us to consider the original implementation of (+++). When he first introduced +++ to us, this was the code he wrote, which I am going to call the original implementation: infixr 5 +++ (+++) :: Parser a -> Parser a -> Parser a p +++ q = Parser $ \s -> runParser p s ++ runParser q s I have been having tons of trouble since we were introduced to parsing and so it continues. I have tried/am considering two approaches. 1) Use the "original" implementation, as in p +++ q = Parser $ \s - runParser p s ++ runParser q s 2) Use the final implementation, as in (+++) = mplus Here are my questions: 1) The module will not compile if I use the original implementation. The error: Not in scope: data constructor 'Parser'. It compiles fine using (+++) = mplus. What is wrong with using the original implementation that is avoided by using the final implementation? 2) How do I check if the first Parser returns anything? Is something like (not (isNothing (Parser $ \s - runParser p s) on the right track? It seems like it should be easy but I have no idea. 3) Once I figure out how to check if the first Parser returns anything, if I am to base my code on the final implementation, would it be as easy as this?: -- if p returns something then p <++ q = mplus (Parser $ \s -> runParser p s) mzero -- else (<++) = mplus Best, Jeff

  Read the article

• Does protobuf-net generated binary compatible with Google specs

  - by cornerback84

  Actually I want to serialize my data using Google's java implementation and then deserialize using C# implementation? I have chosen portobuf-net as it seems to be more stable (porto# is still v0.9 or I would have gone for it). Before I start working on it I wanted to be sure that I can achieve this (serializing data using java implementation and deserializing it using potobuf-net). Or is there any list of methods that are specific to portobuf-net implementation?

  Read the article

• Does portobuf-net generated binary compatible with Google specs

  - by cornerback84

  Actually I want to serialize my data using Google's java implementation and then deserialize using C# implementation? I have chosen portobuf-net as it seems to be more stable (porto# is still v0.9 or I would have gone for it). Before I start working on it I wanted to be sure that I can achieve this (serializing data using java implementation and deserializing it using potobuf-net). Or is there any list of methods that are specific to portobuf-net implementation?

  Read the article

• Usage of @specialized in traits

  - by paradigmatic

  I have a trait and an implementation looking like: trait Foo[A] { def bar[B >: A: Ordering]: Foo[B] } class FooImpl[A]( val a: A, val values: List[Foo[A]] ) extends Foo[A] { def bar[B >: A] = { /* concrete implementation */} } I would like to use the @specialized annotation on A and B to avoid autoboxing. Do I need to use it in both trait and implementation, only in implementation, or only in trait ?

  Read the article

• Is portobuf-net generated binary compatible with Google specs

  - by cornerback84

  Actually I want to serialize my data using Google's java implementation and then deserialize using C# implementation? I have chosen portobuf-net as it seems to be more stable (porto# is still v0.9 or I would have gone for it). Before I start working on it I wanted to be sure that I can achieve this (serializing data using java implementation and deserializing it using potobuf-net). Or is there any list of methods that are specific to portobuf-net implementation?

  Read the article

• Loosely coupled .NET Cache Provider using Dependency Injection

  - by Rhames

  I have recently been reading the excellent book “Dependency Injection in .NET”, written by Mark Seemann. I do not generally buy software development related books, as I never seem to have the time to read them, but I have found the time to read Mark’s book, and it was time well spent I think. Reading the ideas around Dependency Injection made me realise that the Cache Provider code I wrote about earlier (see http://geekswithblogs.net/Rhames/archive/2011/01/10/using-the-asp.net-cache-to-cache-data-in-a-model.aspx) could be refactored to use Dependency Injection, which should produce cleaner code. The goals are to: Separate the cache provider implementation (using the ASP.NET data cache) from the consumers (loose coupling). This will also mean that the dependency on System.Web for the cache provider does not ripple down into the layers where it is being consumed (such as the domain layer). Provide a decorator pattern to allow a consumer of the cache provider to be implemented separately from the base consumer (i.e. if we have a base repository, we can decorate this with a caching version). Although I used the term repository, in reality the cache consumer could be just about anything. Use constructor injection to provide the Dependency Injection, with a suitable DI container (I use Castle Windsor). The sample code for this post is available on github, https://github.com/RobinHames/CacheProvider.git ICacheProvider In the sample code, the key interface is ICacheProvider, which is in the domain layer. 1: using System; 2: using System.Collections.Generic; 3:   4: namespace CacheDiSample.Domain 5: { 6: public interface ICacheProvider<T> 7: { 8: T Fetch(string key, Func<T> retrieveData, DateTime? absoluteExpiry, TimeSpan? relativeExpiry); 9: IEnumerable<T> Fetch(string key, Func<IEnumerable<T>> retrieveData, DateTime? absoluteExpiry, TimeSpan? relativeExpiry); 10: } 11: }   This interface contains two methods to retrieve data from the cache, either as a single instance or as an IEnumerable. the second paramerter is of type Func<T>. This is the method used to retrieve data if nothing is found in the cache. The ASP.NET implementation of the ICacheProvider interface needs to live in a project that has a reference to system.web, typically this will be the root UI project, or it could be a separate project. The key thing is that the domain or data access layers do not need system.web references adding to them. In my sample MVC application, the CacheProvider is implemented in the UI project, in a folder called “CacheProviders”: 1: using System; 2: using System.Collections.Generic; 3: using System.Linq; 4: using System.Web; 5: using System.Web.Caching; 6: using CacheDiSample.Domain; 7:   8: namespace CacheDiSample.CacheProvider 9: { 10: public class CacheProvider<T> : ICacheProvider<T> 11: { 12: public T Fetch(string key, Func<T> retrieveData, DateTime? absoluteExpiry, TimeSpan? relativeExpiry) 13: { 14: return FetchAndCache<T>(key, retrieveData, absoluteExpiry, relativeExpiry); 15: } 16:   17: public IEnumerable<T> Fetch(string key, Func<IEnumerable<T>> retrieveData, DateTime? absoluteExpiry, TimeSpan? relativeExpiry) 18: { 19: return FetchAndCache<IEnumerable<T>>(key, retrieveData, absoluteExpiry, relativeExpiry); 20: } 21:   22: #region Helper Methods 23:   24: private U FetchAndCache<U>(string key, Func<U> retrieveData, DateTime? absoluteExpiry, TimeSpan? relativeExpiry) 25: { 26: U value; 27: if (!TryGetValue<U>(key, out value)) 28: { 29: value = retrieveData(); 30: if (!absoluteExpiry.HasValue) 31: absoluteExpiry = Cache.NoAbsoluteExpiration; 32:   33: if (!relativeExpiry.HasValue) 34: relativeExpiry = Cache.NoSlidingExpiration; 35:   36: HttpContext.Current.Cache.Insert(key, value, null, absoluteExpiry.Value, relativeExpiry.Value); 37: } 38: return value; 39: } 40:   41: private bool TryGetValue<U>(string key, out U value) 42: { 43: object cachedValue = HttpContext.Current.Cache.Get(key); 44: if (cachedValue == null) 45: { 46: value = default(U); 47: return false; 48: } 49: else 50: { 51: try 52: { 53: value = (U)cachedValue; 54: return true; 55: } 56: catch 57: { 58: value = default(U); 59: return false; 60: } 61: } 62: } 63:   64: #endregion 65:   66: } 67: }   The FetchAndCache helper method checks if the specified cache key exists, if it does not, the Func<U> retrieveData method is called, and the results are added to the cache. Using Castle Windsor to register the cache provider In the MVC UI project (my application root), Castle Windsor is used to register the CacheProvider implementation, using a Windsor Installer: 1: using Castle.MicroKernel.Registration; 2: using Castle.MicroKernel.SubSystems.Configuration; 3: using Castle.Windsor; 4:   5: using CacheDiSample.Domain; 6: using CacheDiSample.CacheProvider; 7:   8: namespace CacheDiSample.WindsorInstallers 9: { 10: public class CacheInstaller : IWindsorInstaller 11: { 12: public void Install(IWindsorContainer container, IConfigurationStore store) 13: { 14: container.Register( 15: Component.For(typeof(ICacheProvider<>)) 16: .ImplementedBy(typeof(CacheProvider<>)) 17: .LifestyleTransient()); 18: } 19: } 20: }   Note that the cache provider is registered as a open generic type. Consuming a Repository I have an existing couple of repository interfaces defined in my domain layer: IRepository.cs 1: using System; 2: using System.Collections.Generic; 3:   4: using CacheDiSample.Domain.Model; 5:   6: namespace CacheDiSample.Domain.Repositories 7: { 8: public interface IRepository<T> 9: where T : EntityBase 10: { 11: T GetById(int id); 12: IList<T> GetAll(); 13: } 14: }   IBlogRepository.cs 1: using System; 2: using CacheDiSample.Domain.Model; 3:   4: namespace CacheDiSample.Domain.Repositories 5: { 6: public interface IBlogRepository : IRepository<Blog> 7: { 8: Blog GetByName(string name); 9: } 10: }   These two repositories are implemented in the DataAccess layer, using Entity Framework to retrieve data (this is not important though). One important point is that in the BaseRepository implementation of IRepository, the methods are virtual. This will allow the decorator to override them. The BlogRepository is registered in a RepositoriesInstaller, again in the MVC UI project. 1: using Castle.MicroKernel.Registration; 2: using Castle.MicroKernel.SubSystems.Configuration; 3: using Castle.Windsor; 4:   5: using CacheDiSample.Domain.CacheDecorators; 6: using CacheDiSample.Domain.Repositories; 7: using CacheDiSample.DataAccess; 8:   9: namespace CacheDiSample.WindsorInstallers 10: { 11: public class RepositoriesInstaller : IWindsorInstaller 12: { 13: public void Install(IWindsorContainer container, IConfigurationStore store) 14: { 15: container.Register(Component.For<IBlogRepository>() 16: .ImplementedBy<BlogRepository>() 17: .LifestyleTransient() 18: .DependsOn(new 19: { 20: nameOrConnectionString = "BloggingContext" 21: })); 22: } 23: } 24: }   Now I can inject a dependency on the IBlogRepository into a consumer, such as a controller in my sample code: 1: using System; 2: using System.Collections.Generic; 3: using System.Linq; 4: using System.Web; 5: using System.Web.Mvc; 6:   7: using CacheDiSample.Domain.Repositories; 8: using CacheDiSample.Domain.Model; 9:   10: namespace CacheDiSample.Controllers 11: { 12: public class HomeController : Controller 13: { 14: private readonly IBlogRepository blogRepository; 15:   16: public HomeController(IBlogRepository blogRepository) 17: { 18: if (blogRepository == null) 19: throw new ArgumentNullException("blogRepository"); 20:   21: this.blogRepository = blogRepository; 22: } 23:   24: public ActionResult Index() 25: { 26: ViewBag.Message = "Welcome to ASP.NET MVC!"; 27:   28: var blogs = blogRepository.GetAll(); 29:   30: return View(new Models.HomeModel { Blogs = blogs }); 31: } 32:   33: public ActionResult About() 34: { 35: return View(); 36: } 37: } 38: }   Consuming the Cache Provider via a Decorator I used a Decorator pattern to consume the cache provider, this means my repositories follow the open/closed principle, as they do not require any modifications to implement the caching. It also means that my controllers do not have any knowledge of the caching taking place, as the DI container will simply inject the decorator instead of the root implementation of the repository. The first step is to implement a BlogRepository decorator, with the caching logic in it. Note that this can reside in the domain layer, as it does not require any knowledge of the data access methods. BlogRepositoryWithCaching.cs 1: using System; 2: using System.Collections.Generic; 3: using System.Linq; 4: using System.Text; 5:   6: using CacheDiSample.Domain.Model; 7: using CacheDiSample.Domain; 8: using CacheDiSample.Domain.Repositories; 9:   10: namespace CacheDiSample.Domain.CacheDecorators 11: { 12: public class BlogRepositoryWithCaching : IBlogRepository 13: { 14: // The generic cache provider, injected by DI 15: private ICacheProvider<Blog> cacheProvider; 16: // The decorated blog repository, injected by DI 17: private IBlogRepository parentBlogRepository; 18:   19: public BlogRepositoryWithCaching(IBlogRepository parentBlogRepository, ICacheProvider<Blog> cacheProvider) 20: { 21: if (parentBlogRepository == null) 22: throw new ArgumentNullException("parentBlogRepository"); 23:   24: this.parentBlogRepository = parentBlogRepository; 25:   26: if (cacheProvider == null) 27: throw new ArgumentNullException("cacheProvider"); 28:   29: this.cacheProvider = cacheProvider; 30: } 31:   32: public Blog GetByName(string name) 33: { 34: string key = string.Format("CacheDiSample.DataAccess.GetByName.{0}", name); 35: // hard code 5 minute expiry! 36: TimeSpan relativeCacheExpiry = new TimeSpan(0, 5, 0); 37: return cacheProvider.Fetch(key, () => 38: { 39: return parentBlogRepository.GetByName(name); 40: }, 41: null, relativeCacheExpiry); 42: } 43:   44: public Blog GetById(int id) 45: { 46: string key = string.Format("CacheDiSample.DataAccess.GetById.{0}", id); 47:   48: // hard code 5 minute expiry! 49: TimeSpan relativeCacheExpiry = new TimeSpan(0, 5, 0); 50: return cacheProvider.Fetch(key, () => 51: { 52: return parentBlogRepository.GetById(id); 53: }, 54: null, relativeCacheExpiry); 55: } 56:   57: public IList<Blog> GetAll() 58: { 59: string key = string.Format("CacheDiSample.DataAccess.GetAll"); 60:   61: // hard code 5 minute expiry! 62: TimeSpan relativeCacheExpiry = new TimeSpan(0, 5, 0); 63: return cacheProvider.Fetch(key, () => 64: { 65: return parentBlogRepository.GetAll(); 66: }, 67: null, relativeCacheExpiry) 68: .ToList(); 69: } 70: } 71: }   The key things in this caching repository are: I inject into the repository the ICacheProvider<Blog> implementation, via the constructor. This will make the cache provider functionality available to the repository. I inject the parent IBlogRepository implementation (which has the actual data access code), via the constructor. This will allow the methods implemented in the parent to be called if nothing is found in the cache. I override each of the methods implemented in the repository, including those implemented in the generic BaseRepository. Each override of these methods follows the same pattern. It makes a call to the CacheProvider.Fetch method, and passes in the parentBlogRepository implementation of the method as the retrieval method, to be used if nothing is present in the cache. Configuring the Caching Repository in the DI Container The final piece of the jigsaw is to tell Castle Windsor to use the BlogRepositoryWithCaching implementation of IBlogRepository, but to inject the actual Data Access implementation into this decorator. This is easily achieved by modifying the RepositoriesInstaller to use Windsor’s implicit decorator wiring: 1: using Castle.MicroKernel.Registration; 2: using Castle.MicroKernel.SubSystems.Configuration; 3: using Castle.Windsor; 4:   5: using CacheDiSample.Domain.CacheDecorators; 6: using CacheDiSample.Domain.Repositories; 7: using CacheDiSample.DataAccess; 8:   9: namespace CacheDiSample.WindsorInstallers 10: { 11: public class RepositoriesInstaller : IWindsorInstaller 12: { 13: public void Install(IWindsorContainer container, IConfigurationStore store) 14: { 15:   16: // Use Castle Windsor implicit wiring for the block repository decorator 17: // Register the outermost decorator first 18: container.Register(Component.For<IBlogRepository>() 19: .ImplementedBy<BlogRepositoryWithCaching>() 20: .LifestyleTransient()); 21: // Next register the IBlogRepository inmplementation to inject into the outer decorator 22: container.Register(Component.For<IBlogRepository>() 23: .ImplementedBy<BlogRepository>() 24: .LifestyleTransient() 25: .DependsOn(new 26: { 27: nameOrConnectionString = "BloggingContext" 28: })); 29: } 30: } 31: }   This is all that is needed. Now if the consumer of the repository makes a call to the repositories method, it will be routed via the caching mechanism. You can test this by stepping through the code, and seeing that the DataAccess.BlogRepository code is only called if there is no data in the cache, or this has expired. The next step is to add the SQL Cache Dependency support into this pattern, this will be a future post.

  Read the article

• Scope quandary with namespaces, function templates, and static data

  - by Adrian McCarthy

  This scoping problem seems like the type of C++ quandary that Scott Meyers would have addressed in one of his Effective C++ books. I have a function, Analyze, that does some analysis on a range of data. The function is called from a few places with different types of iterators, so I have made it a template (and thus implemented it in a header file). The function depends on a static table of data, AnalysisTable, that I don't want to expose to the rest of the code. My first approach was to make the table a static const inside Analysis. namespace MyNamespace { template <typename InputIterator> int Analyze(InputIterator begin, InputIterator end) { static const int AnalysisTable[] = { /* data */ }; ... // implementation uses AnalysisTable return result; } } // namespace MyNamespace It appears that the compiler creates a copy of AnalysisTable for each instantiation of Analyze, which is wasteful of space (and, to a small degree, time). So I moved the table outside the function like this: namespace MyNamespace { const int AnalysisTable[] = { /* data */ }; template <typename InputIterator> int Analyze(InputIterator begin, InputIterator end) { ... // implementation uses AnalysisTable return result; } } // namespace MyNamespace There's only one copy of the table now, but it's exposed to the rest of the code. I'd rather keep this implementation detail hidden, so I introduced an unnamed namespace: namespace MyNamespace { namespace { // unnamed to hide AnalysisTable const int AnalysisTable[] = { /* data */ }; } // unnamed namespace template <typename InputIterator> int Analyze(InputIterator begin, InputIterator end) { ... // implementation uses AnalysisTable return result; } } // namespace MyNamespace But now I again have multiple copies of the table, because each compilation unit that includes this header file gets its own. If Analyze weren't a template, I could move all the implementation detail out of the header file. But it is a template, so I seem stuck. My next attempt was to put the table in the implementation file and to make an extern declaration within Analyze. // foo.h ------ namespace MyNamespace { template <typename InputIterator> int Analyze(InputIterator begin, InputIterator end) { extern const int AnalysisTable[]; ... // implementation uses AnalysisTable return result; } } // namespace MyNamespace // foo.cpp ------ #include "foo.h" namespace MyNamespace { const int AnalysisTable[] = { /* data */ }; } This looks like it should work, and--indeed--the compiler is satisfied. The linker, however, complains, "unresolved external symbol AnalysisTable." Drat! (Can someone explain what I'm missing here?) The only thing I could think of was to give the inner namespace a name, declare the table in the header, and provide the actual data in an implementation file: // foo.h ----- namespace MyNamespace { namespace PrivateStuff { extern const int AnalysisTable[]; } // unnamed namespace template <typename InputIterator> int Analyze(InputIterator begin, InputIterator end) { ... // implementation uses PrivateStuff::AnalysisTable return result; } } // namespace MyNamespace // foo.cpp ----- #include "foo.h" namespace MyNamespace { namespace PrivateStuff { const int AnalysisTable[] = { /* data */ }; } } Once again, I have exactly one instance of AnalysisTable (yay!), but other parts of the program can access it (boo!). The inner namespace makes it a little clearer that they shouldn't, but it's still possible. Is it possible to have one instance of the table and to move the table beyond the reach of everything but Analyze?

  Read the article

• Difference between performSelectorInBackground and NSOperation Subclass

  - by AmitSri

  I have created one testing app for running deep counter loop. I run the loop fuction in background thread using performSelectorInBackground and also NSOperation subclass separately. I am also using performSelectorOnMainThread to notify main thread within backgroundthread method and [NSNotificationCenter defaultCenter] postNotificationName within NSOperation subclass to notify main thread for updating UI. Initially both the implementation giving me same result and i am able to update UI without having any problem. The only difference i found is the Thread count between two implementations. The performSelectorInBackground implementation created one thread and got terminated after loop finished and my app thread count again goes to 1. The NSOperation subclass implementation created two new threads and keep exists in the application and i can see 3 threads after loop got finished in main() function. So, my question is why two threads created by NSOperation and why it didn't get terminated just like the first background thread implementation? I am little bit confuse and unable to decide which implementation is best in-terms of performance and memory management. Thanks

  Read the article

• Better to build or buy a compute grid platform?

  - by James B

  I am looking to do some quite processor-intensive brute force processing for string matching. I have run my prototype in a multi-threaded environment and compared the performance to an implementation using Gridgain with a couple of nodes (also multithreaded). The performance I observed was that my Gridgain implementation performed slower to my multithreaded implementation. It could be the case that there was a flaw in my gridgain implementation, but it was only a prototype, and I thought the results were indicative. So my question is this: What are the advantages of having to learn and then build an implementation for a particular grid platform (hadoop, gridgain, or EC2 if going hosted - other suggestions welcome), when one could fairly easily put together a lightweight compute grid platform with a much shallower learning curve?...i.e. what do we get for free with these cloud/grid platforms that are worth having/tricky to implement? (Please note, I don't have any need for a data grid) Cheers, -James (p.s. Happy to make this community wiki if needbe)

  Read the article

• Implementing a Custom Coherence PartitionAssignmentStrategy

  - by jpurdy

  A recent A-Team engagement required the development of a custom PartitionAssignmentStrategy (PAS). By way of background, a PAS is an implementation of a Java interface that controls how a Coherence partitioned cache service assigns partitions (primary and backup copies) across the available set of storage-enabled members. While seemingly straightforward, this is actually a very difficult problem to solve. Traditionally, Coherence used a distributed algorithm spread across the cache servers (and as of Coherence 3.7, this is still the default implementation). With the introduction of the PAS interface, the model of operation was changed so that the logic would run solely in the cache service senior member. Obviously, this makes the development of a custom PAS vastly less complex, and in practice does not introduce a significant single point of failure/bottleneck. Note that Coherence ships with a default PAS implementation but it is not used by default. Further, custom PAS implementations are uncommon (this engagement was the first custom implementation that we know of). The particular implementation mentioned above also faced challenges related to managing multiple backup copies but that won't be discussed here. There were a few challenges that arose during design and implementation: Naive algorithms had an unreasonable upper bound of computational cost. There was significant complexity associated with configurations where the member count varied significantly between physical machines. Most of the complexity of a PAS is related to rebalancing, not initial assignment (which is usually fairly simple). A custom PAS may need to solve several problems simultaneously, such as: Ensuring that each member has a similar number of primary and backup partitions (e.g. each member has the same number of primary and backup partitions) Ensuring that each member carries similar responsibility (e.g. the most heavily loaded member has no more than one partition more than the least loaded). Ensuring that each partition is on the same member as a corresponding local resource (e.g. for applications that use partitioning across message queues, to ensure that each partition is collocated with its corresponding message queue). Ensuring that a given member holds no more than a given number of partitions (e.g. no member has more than 10 partitions) Ensuring that backups are placed far enough away from the primaries (e.g. on a different physical machine or a different blade enclosure) Achieving the above goals while ensuring that partition movement is minimized. These objectives can be even more complicated when the topology of the cluster is irregular. For example, if multiple cluster members may exist on each physical machine, then clearly the possibility exists that at certain points (e.g. following a member failure), the number of members on each machine may vary, in certain cases significantly so. Consider the case where there are three physical machines, with 3, 3 and 9 members each (respectively). This introduces complexity since the backups for the 9 members on the the largest machine must be spread across the other 6 members (to ensure placement on different physical machines), preventing an even distribution. For any given problem like this, there are usually reasonable compromises available, but the key point is that objectives may conflict under extreme (but not at all unlikely) circumstances. The most obvious general purpose partition assignment algorithm (possibly the only general purpose one) is to define a scoring function for a given mapping of partitions to members, and then apply that function to each possible permutation, selecting the most optimal permutation. This would result in N! (factorial) evaluations of the scoring function. This is clearly impractical for all but the smallest values of N (e.g. a partition count in the single digits). It's difficult to prove that more efficient general purpose algorithms don't exist, but the key take away from this is that algorithms will tend to either have exorbitant worst case performance or may fail to find optimal solutions (or both) -- it is very important to be able to show that worst case performance is acceptable. This quickly leads to the conclusion that the problem must be further constrained, perhaps by limiting functionality or by using domain-specific optimizations. Unfortunately, it can be very difficult to design these more focused algorithms. In the specific case mentioned, we constrained the solution space to very small clusters (in terms of machine count) with small partition counts and supported exactly two backup copies, and accepted the fact that partition movement could potentially be significant (preferring to solve that issue through brute force). We then used the out-of-the-box PAS implementation as a fallback, delegating to it for configurations that were not supported by our algorithm. Our experience was that the PAS interface is quite usable, but there are intrinsic challenges to designing PAS implementations that should be very carefully evaluated before committing to that approach.

  Read the article

• Merge sort versus quick sort performance

  - by Giorgio

  I have implemented merge sort and quick sort using C (GCC 4.4.3 on Ubuntu 10.04 running on a 4 GB RAM laptop with an Intel DUO CPU at 2GHz) and I wanted to compare the performance of the two algorithms. The prototypes of the sorting functions are: void merge_sort(const char **lines, int start, int end); void quick_sort(const char **lines, int start, int end); i.e. both take an array of pointers to strings and sort the elements with index i : start <= i <= end. I have produced some files containing random strings with length on average 4.5 characters. The test files range from 100 lines to 10000000 lines. I was a bit surprised by the results because, even though I know that merge sort has complexity O(n log(n)) while quick sort is O(n^2), I have often read that on average quick sort should be as fast as merge sort. However, my results are the following. Up to 10000 strings, both algorithms perform equally well. For 10000 strings, both require about 0.007 seconds. For 100000 strings, merge sort is slightly faster with 0.095 s against 0.121 s. For 1000000 strings merge sort takes 1.287 s against 5.233 s of quick sort. For 5000000 strings merge sort takes 7.582 s against 118.240 s of quick sort. For 10000000 strings merge sort takes 16.305 s against 1202.918 s of quick sort. So my question is: are my results as expected, meaning that quick sort is comparable in speed to merge sort for small inputs but, as the size of the input data grows, the fact that its complexity is quadratic will become evident? Here is a sketch of what I did. In the merge sort implementation, the partitioning consists in calling merge sort recursively, i.e. merge_sort(lines, start, (start + end) / 2); merge_sort(lines, 1 + (start + end) / 2, end); Merging of the two sorted sub-array is performed by reading the data from the array lines and writing it to a global temporary array of pointers (this global array is allocate only once). After each merge the pointers are copied back to the original array. So the strings are stored once but I need twice as much memory for the pointers. For quick sort, the partition function chooses the last element of the array to sort as the pivot and scans the previous elements in one loop. After it has produced a partition of the type start ... {elements <= pivot} ... pivotIndex ... {elements > pivot} ... end it calls itself recursively: quick_sort(lines, start, pivotIndex - 1); quick_sort(lines, pivotIndex + 1, end); Note that this quick sort implementation sorts the array in-place and does not require additional memory, therefore it is more memory efficient than the merge sort implementation. So my question is: is there a better way to implement quick sort that is worthwhile trying out? If I improve the quick sort implementation and perform more tests on different data sets (computing the average of the running times on different data sets) can I expect a better performance of quick sort wrt merge sort? EDIT Thank you for your answers. My implementation is in-place and is based on the pseudo-code I have found on wikipedia in Section In-place version: function partition(array, 'left', 'right', 'pivotIndex') where I choose the last element in the range to be sorted as a pivot, i.e. pivotIndex := right. I have checked the code over and over again and it seems correct to me. In order to rule out the case that I am using the wrong implementation I have uploaded the source code on github (in case you would like to take a look at it). Your answers seem to suggest that I am using the wrong test data. I will look into it and try out different test data sets. I will report as soon as I have some results.

  Read the article

• Iterative and Incremental Principle Series 4: Iteration Planning – (a.k.a What should I do today?)

  - by llowitz

  Welcome back to the fourth of a five part series on applying the Iteration and Incremental principle.  During the last segment, we discussed how the Implementation Plan includes the number of the iterations for a project, but not the specifics about what will occur during each iteration.  Today, we will explore Iteration Planning and discuss how and when to plan your iterations. As mentioned yesterday, OUM prescribes initially planning your project approach at a high level by creating an Implementation Plan.  As the project moves through the lifecycle, the plan is progressively refined.  Specifically, the details of each iteration is planned prior to the iteration start. The Iteration Plan starts by identifying the iteration goal.  An example of an iteration goal during the OUM Elaboration Phase may be to complete the RD.140.2 Create Requirements Specification for a specific set of requirements.  Another project may determine that their iteration goal is to focus on a smaller set of requirements, but to complete both the RD.140.2 Create Requirements Specification and the AN.100.1 Prepare Analysis Specification.  In an OUM project, the Iteration Plan needs to identify both the iteration goal – how far along the implementation lifecycle you plan to be, and the scope of work for the iteration.  Since each iteration typically ranges from 2 weeks to 6 weeks, it is important to identify a scope of work that is achievable, yet challenging, given the iteration goal and timeframe.  OUM provides specific guidelines and techniques to help prioritize the scope of work based on criteria such as risk, complexity, customer priority and dependency.  In OUM, this prioritization helps focus early iterations on the high risk, architecturally significant items helping to mitigate overall project risk.  Central to the prioritization is the MoSCoW (Must Have, Should Have, Could Have, and Won’t Have) list.   The result of the MoSCoW prioritization is an Iteration Group.  This is a scope of work to be worked on as a group during one or more iterations.  As I mentioned during yesterday’s blog, it is pointless to plan my daily exercise in advance since several factors, including the weather, influence what exercise I perform each day.  Therefore, every morning I perform Iteration Planning.   My “Iteration Plan” includes the type of exercise for the day (run, bike, elliptical), whether I will exercise outside or at the gym, and how many interval sets I plan to complete.    I use several factors to prioritize the type of exercise that I perform each day.  Since running outside is my highest priority, I try to complete it early in the week to minimize the risk of not meeting my overall goal of doing it twice each week.  Regardless of the specific exercise I select, I follow the guidelines in my Implementation Plan by applying the 6-minute interval sets.  Just as in OUM, the iteration goal should be in context of the overall Implementation Plan, and the iteration goal should move the project closer to achieving the phase milestone goals. Having an Implementation Plan details the strategy of what I plan to do and keeps me on track, while the Iteration Plan affords me the flexibility to juggle what I do each day based on external influences thus maximizing my overall success. Tomorrow I’ll conclude the series on applying the Iterative and Incremental approach by discussing how to manage the iteration duration and highlighting some benefits of applying this principle.

  Read the article

• CI tests to enforce specific development rules - good practice?

  - by KeithS

  The following is all purely hypothetical and any particular portion of it may or may not accurately describe real persons or situations, whether living, dead or just pretending. Let's say I'm a senior dev or architect in charge of a dev team working on a project. This project includes a security library for user authentication/authorization of the application under development. The library must be available for developers to edit; however, I wish to "trust but verify" that coders are not doing things that could compromise the security of the finished system, and because this isn't my only responsibility I want it to be done in an automated way. As one example, let's say I have an interface that represents a user which has been authenticated by the system's security library. The interface exposes basic user info and a list of things the user is authorized to do (so that the client app doesn't have to keep asking the server "can I do this?"), all in an immutable fashion of course. There is only one implementation of this interface in production code, and for the purposes of this post we can say that all appropriate measures have been taken to ensure that this implementation can only be used by the one part of our code that needs to be able to create concretions of the interface. The coders have been instructed that this interface and its implementation are sacrosanct and any changes must go through me. However, those are just words; the security library's source is open for editing by necessity. Any of my devs could decide that this secured, private, hash-checked implementation needs to be public so that they could do X, or alternately they could create their own implementation of this public interface in a different library, exposing the hashing algorithm that provides the secure checksum, in order to do Y. I may not be made aware of these changes so that I can beat the developer over the head for it. An attacker could then find these little nuggets in an unobfuscated library of the compiled product, and exploit it to provide fake users and/or falsely-elevated administrative permissions, bypassing the entire security system. This possibility keeps me awake for a couple of nights, and then I create an automated test that reflectively checks the codebase for types deriving from the interface, and fails if it finds any that are not exactly what and where I expect them to be. I compile this test into a project under a separate folder of the VCS that only I have rights to commit to, have CI compile it as an external library of the main project, and set it up to run as part of the CI test suite for user commits. Now, I have an automated test under my complete control that will tell me (and everyone else) if the number of implementations increases without my involvement, or an implementation that I did know about has anything new added or has its modifiers or those of its members changed. I can then investigate further, and regain the opportunity to beat developers over the head as necessary. Is this considered "reasonable" to want to do in situations like this? Am I going to be seen in a negative light for going behind my devs' backs to ensure they aren't doing something they shouldn't?

  Read the article

• How to model has_many with polymorphism?

  - by Daniel Abrahamsson

  I've run into a situation that I am not quite sure how to model. Suppose I have a User class, and a user has many services. However, these services are quite different, for example a MailService and a BackupService, so single table inheritance won't do. Instead, I am thinking of using polymorphic associations together with an abstract base class: class User < ActiveRecord::Base has_many :services end class Service < ActiveRecord::Base validates_presence_of :user_id, :implementation_id, :implementation_type belongs_to :user belongs_to :implementation, :polymorphic = true delegate :common_service_method, :name, :to => :implementation end #Base class for service implementations class ServiceImplementation < ActiveRecord::Base validates_presence_of :user_id, :on => :create has_one :service, :as => :implementation has_one :user, :through => :service after_create :create_service_record #Tell Rails this class does not use a table. def self.abstract_class? true end #Default name implementation. def name self.class.name end protected #Sets up a service object def create_service_record service = Service.new(:user_id => user_id) service.implementation = self service.save! end end class MailService < ServiceImplementation #validations, etc... def common_service_method puts "MailService implementation of common service method" end end #Example usage MailService.create(..., :user_id => user.id) BackupService.create(...., :user_id => user.id) user.services.each do |s| puts "#{user.name} is using #{s.name}" end #Daniel is using MailService, Daniel is using BackupService So, is this the best solution? Or even a good one? How have you solved this kind of problem?

  Read the article

• Is there a way to increase the efficiency of shared_ptr by storing the reference count inside the co

  - by BillyONeal

  Hello everyone :) This is becoming a common pattern in my code, for when I need to manage an object that needs to be noncopyable because either A. it is "heavy" or B. it is an operating system resource, such as a critical section: class Resource; class Implementation : public boost::noncopyable { friend class Resource; HANDLE someData; Implementation(HANDLE input) : someData(input) {}; void SomeMethodThatActsOnHandle() { //Do stuff }; public: ~Implementation() { FreeHandle(someData) }; }; class Resource { boost::shared_ptr<Implementation> impl; public: Resource(int argA) explicit { HANDLE handle = SomeLegacyCApiThatMakesSomething(argA); if (handle == INVALID_HANDLE_VALUE) throw SomeTypeOfException(); impl.reset(new Implementation(handle)); }; void SomeMethodThatActsOnTheResource() { impl->SomeMethodThatActsOnTheHandle(); }; }; This way, shared_ptr takes care of the reference counting headaches, allowing Resource to be copyable, even though the underlying handle should only be closed once all references to it are destroyed. However, it seems like we could save the overhead of allocating shared_ptr's reference counts and such separately if we could move that data inside Implementation somehow, like boost's intrusive containers do. If this is making the premature optimization hackles nag some people, I actually agree that I don't need this for my current project. But I'm curious if it is possible.

  Read the article

• Documentation in Oracle Retail Merchandising System (RMS) and Oracle Retail Fiscal Management System (ORFM), Release 13.2.4

  - by Oracle Retail Documentation Team

  The Patch Release 13.2.4 of the Oracle Retail Merchandising System (RMS) and its module, Oracle Retail Fiscal Management (ORFM)  is now available from My Oracle Support. End User Documentation Enhancements The following summarize the highlights of changes made to the documentation in conjunction with the new Brazil-related functionality: Foundation chapter in the Oracle Retail Merchandising System (RMS)/Sales Audit (ReSA) Brazil Localization User GuideThis chapter was updated with a non-base Localization Flexible Attribution Solution (LFAS) section that addresses the addition of several new custom attributes to Items and Suppliers through non-base LFAS for Brazil; it also addresses the extension of the Retail Tax Integration Layer (RTIL) through the Oracle Retail Merchandising System (RMS), and Oracle Retail Fiscal Management System (ORFM).  ORFM User GuideThe Purchase Order chapter was updated to include schedule related updates for a Nota Fiscal. The Fiscal Documents chapter was updated to include information on creating a new NF and searching for details using Vendor Product Number. Oracle Retail Fiscal Management/RMS Brazil Localization Implementation GuideThe Implementation Checklist chapter was updated with a note on multi-currency functionality. The Batch Processes chapter was updated with information on the NF EDI batch. The following summarize the highlights of changes made to the documentation in conjunction with the new technical certifications (see the RMS 13.2.4 Release Notes for more information): Installation Guides for RMS and for ORFM/RMS BrazilThese installation guides were updated extensively to account for the multiple technical certification enhancements in 13.2.4. White Paper: How to Upgrade from WebLogic11g 10.3.3 to WebLogic11g 10.3.4  (Doc ID: 1432575.1)See the previous blog entry regarding this new White Paper. New Documents on My Oracle Support for Brazil Localization Overview and Interfaces Tax Vendor Integration (Doc ID: 1424048.1)Oracle chooses to integrate with a third party tax expert to delivery the Brazilian solution. Oracle has built the Retail Tax Integration layer (RTIL) as the key integration component to support the integration of Oracle suite of products with external tax vendors. This paper addresses the RTIL integration interfaces with TaxWeb, providing guidance on the typical integration interfaces and operations that must be supported by other tax solutions in the Brazilian market. Oracle Retail Fiscal Management/RMS Brazil Localization: Localization Flexible Attribute Solution (LFAS) (Doc ID: 1418509.1)The white paper covers the definition of custom attributes in Localization Flexible Attribute Solution (LFAS) and enables retailers to perform data conversion changes. Retailers can add several new custom attributes to Items and Suppliers through non-base LFAS for Brazil and extend Retail Tax Integration Layer (RTIL) through the Oracle Retail Merchandising System (RMS), and Oracle Retail Fiscal Management System (RFM). Documents Published in RMS and ORFM Release 13.2.4 Oracle Retail Merchandising System Release Notes Oracle Retail Merchandising System Installation Guide Oracle Retail Merchandising System User Guide and Online Help Oracle Retail Sales Audit (ReSA) User Guide and Online Help Oracle Retail Merchandising System Operations Guide Oracle Retail Merchandising System Data Model Oracle Retail Merchandising Batch Schedule Oracle Retail Merchandising Implementation Guide Oracle Retail POS Suite 13.4.1 / Merchandising Operations Management13.2.4 Implementation Guide Oracle Retail Fiscal Management Data Model Oracle Retail Fiscal Management/RMS Brazil Localization Installation Guide Oracle Retail Fiscal Management/RMS Brazil Localization Implementation Guide Oracle Retail Fiscal Management User Guide and Online Help

  Read the article

• Event Driven Behavior Tree: deterministic traversal order with parallel

  - by Heisenbug

  I've studied several articles and listen some talks about behavior trees (mostly the resources available on AIGameDev by Alex J. Champandard). I'm particularly interested on event driven behavior trees, but I have still some doubts on how to implement them correctly using a scheduler. Just a quick recap: Standard Behavior Tree Each execution tick the tree is traversed from the root in depth-first order The execution order is implicitly expressed by the tree structure. So in the case of behaviors parented to a parallel node, even if both children are executed during the same traversing, the first leaf is always evaluated first. Event Driven BT During the first traversal the nodes (tasks) are enqueued using a scheduler which is responsible for updating only running ones every update The first traversal implicitly produce a depth-first ordered queue in the scheduler Non leaf nodes stays suspended mostly of the time. When a leaf node terminate(either with success or fail status) the parent (observer) is waked up allowing the tree traversing to continue and new tasks will be enqueued in the scheduler Without parallel nodes in the tree there will be up to 1 task running in the scheduler Without parallel nodes, the tasks in the queue(excluding dynamic priority implementation) will be always ordered in a depth-first order (is this right?) Now, from what is my understanding of a possible implementation, there are 2 requirements I think must be respected(I'm not sure though): Now, some requirements I think needs to be guaranteed by a correct implementation are: The result of the traversing should be independent from which implementation strategy is used. The traversing result must be deterministic. I'm struggling trying to guarantee both in the case of parallel nodes. Here's an example: Parallel_1 -->Sequence_1 ---->leaf_A ---->leaf_B -->leaf_C Considering a FIFO policy of the scheduler, before leaf_A node terminates the tasks in the scheduler are: P1(suspended),S1(suspended),leaf_A(running),leaf_C(running) When leaf_A terminate leaf_B will be scheduled (at the end of the queue), so the queue will become: P1(suspended),S1(suspended),leaf_C(running),leaf_B(running) In this case leaf_B will be executed after leaf_C at every update, meanwhile with a non event-driven traversing from the root node, the leaf_B will always be evaluated before leaf_A. So I have a couple of question: do I have understand correctly how event driven BT work? How can I guarantee the depth first order is respected with such an implementation? is this a common issue or am I missing something?

  Read the article

• IXRepository and test problems

  - by Ridermansb

  Recently had a doubt about how and where to test repository methods. Let the following situation: I have an interface IRepository like this: public interface IRepository<T> where T: class, IEntity { IQueryable<T> Query(Expression<Func<T, bool>> expression); // ... Omitted } And a generic implementation of IRepository public class Repository<T> : IRepository<T> where T : class, IEntity { public IQueryable<T> Query(Expression<Func<T, bool>> expression) { return All().Where(expression).AsQueryable(); } } This is an implementation base that can be used by any repository. It contains the basic implementation of my ORM. Some repositories have specific filters, in which case we will IEmployeeRepository with a specific filter: public interface IEmployeeRepository : IRepository<Employee> { IQueryable<Employee> GetInactiveEmployees(); } And the implementation of IEmployeeRepository: public class EmployeeRepository : Repository<Employee>, IEmployeeRepository // TODO: I have a dependency with ORM at this point in Repository<Employee>. How to solve? How to test the GetInactiveEmployees method { public IQueryable<Employee> GetInactiveEmployees() { return Query(p => p.Status != StatusEmployeeEnum.Active || p.StartDate < DateTime.Now); } } Questions Is right to inherit Repository<Employee>? The goal is to reuse code once all implementing IRepository already been made. If EmployeeRepository inherit only IEmployeeRepository, I have to literally copy and paste the code of Repository<T>. In our example, in EmployeeRepository : Repository<Employee> our Repository lies in our ORM layer. We have a dependency here with our ORM impossible to perform some unit test. How to create a unit test to ensure that the filter GetInactiveEmployees return all Employees in which the Status != Active and StartDate < DateTime.Now. I can not create a Fake/Mock of IEmployeeRepository because I would be testing? Need to test the actual implementation of GetInactiveEmployees. The complete code can be found on Github

  Read the article

• UAT Testing for SOA 10G Clusters

  - by [email protected]

  A lot of customers ask how to verify their SOA clusters and make them production ready. Here is a list that I recommend using for 10G SOA Clusters. v\:* {behavior:url(#default#VML);} o\:* {behavior:url(#default#VML);} w\:* {behavior:url(#default#VML);} .shape {behavior:url(#default#VML);} Normal 0 false false false EN-CA X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin-top:0cm; mso-para-margin-right:0cm; mso-para-margin-bottom:10.0pt; mso-para-margin-left:0cm; line-height:115%; mso-pagination:widow-orphan; font-size:11.0pt; mso-bidi-font-size:12.0pt; font-family:"Calibri","sans-serif"; mso-fareast-language:EN-US;} Test cases for each component - Oracle Application Server 10G General Application Server test cases This section is going to cover very General test cases to make sure that the Application Server cluster has been set up correctly and if you can start and stop all the components in the server via opmnct and AS Console. Test Case 1 Check if you can see AS instances in the console Implementation 1. Log on to the AS Console --> check to see if you can see all the nodes in your AS cluster. You should be able to see all the Oracle AS instances that are part of the cluster. This means that the OPMN clustering worked and the AS instances successfully joined the AS cluster. Result You should be able to see if all the instances in the AS cluster are listed in the EM console. If the instances are not listed here are the files to check to see if OPMN joined the cluster properly: $ORACLE_HOME\opmn\logs{*}opmn.log*$ORACLE_HOME\opmn\logs{*}opmn.dbg* If OPMN did not join the cluster properly, please check the opmn.xml file to make sure the discovery multicast address and port are correct (see this link  for opmn documentation). Restart the whole instance using opmnctl stopall followed by opmnctl startall. Log on to AS console to see if instance is listed as part of the cluster. Test Case 2 Check to see if you can start/stop each component Implementation Check each OC4J component on each AS instanceStart each and every component through the AS console to see if they will start and stop.Do that for each and every instance. Result Each component should start and stop through the AS console. You can also verify if the component started by checking opmnctl status by logging onto each box associated with the cluster Test Case 3 Add/modify a datasource entry through AS console on a remote AS instance (not on the instance where EM is physically running) Implementation Pick an OC4J instanceCreate a new data-source through the AS consoleModify an existing data-source or connection pool (optional) Result Open $ORACLE_HOME\j2ee\<oc4j_name>\config\data-sources.xml to see if the new (and or the modified) connection details and data-source exist. If they do then the AS console has successfully updated a remote file and MBeans are communicating correctly. Test Case 4 Start and stop AS instances using opmnctl @cluster command Implementation 1. Go to $ORACLE_HOME\opmn\bin and use the opmnctl @cluster to start and stop the AS instances Result Use opmnctl @cluster status to check for start and stop statuses.  HTTP server test cases This section will deal with use cases to test HTTP server failover scenarios. In these examples the HTTP server will be talking to the BPEL console (or any other web application that the client wants), so the URL will be _http://hostname:port\BPELConsole Test Case 1  Shut down one of the HTTP servers while accessing the BPEL console and see the requested routed to the second HTTP server in the cluster Implementation Access the BPELConsoleCheck $ORACLE_HOME\Apache\Apache\logs\access_log --> check for the timestamp and the URL that was accessed by the user. Timestamp and URL would look like this 1xx.2x.2xx.xxx [24/Mar/2009:16:04:38 -0500] "GET /BPELConsole=System HTTP/1.1" 200 15 After you have figured out which HTTP server this is running on, shut down this HTTP server by using opmnctl stopproc --> this is a graceful shutdown.Access the BPELConsole again (please note that you should have a LoadBalancer in front of the HTTP server and configured the Apache Virtual Host, see EDG for steps)Check $ORACLE_HOME\Apache\Apache\logs\access_log --> check for the timestamp and the URL that was accessed by the user. Timestamp and URL would look like above Result Even though you are shutting down the HTTP server the request is routed to the surviving HTTP server, which is then able to route the request to the BPEL Console and you are able to access the console. By checking the access log file you can confirm that the request is being picked up by the surviving node. Test Case 2 Repeat the same test as above but instead of calling opmnctl stopproc, pull the network cord of one of the HTTP servers, so that the LBR routes the request to the surviving HTTP node --> this is simulating a network failure. Test Case 3 In test case 1 we have simulated a graceful shutdown, in this case we will simulate an Apache crash Implementation Use opmnctl status -l to get the PID of the HTTP server that you would like forcefully bring downOn Linux use kill -9 <PID> to kill the HTTP serverAccess the BPEL console Result As you shut down the HTTP server, OPMN will restart the HTTP server. The restart may be so quick that the LBR may still route the request to the same server. One way to check if the HTTP server restared is to check the new PID and the timestamp in the access log for the BPEL console. BPEL test cases This section is going to cover scenarios dealing with BPEL clustering using jGroups, BPEL deployment and testing related to BPEL failover. Test Case 1 Verify that jGroups has initialized correctly. There is no real testing in this use case just a visual verification by looking at log files that jGroups has initialized correctly. Check the opmn log for the BPEL container for all nodes at $ORACLE_HOME/opmn/logs/<group name><container name><group name>~1.log. This logfile will contain jGroups related information during startup and steady-state operation. Soon after startup you should find log entries for UDP or TCP.Example jGroups Log Entries for UDPApr 3, 2008 6:30:37 PM org.collaxa.thirdparty.jgroups.protocols.UDP createSockets ·         INFO: sockets will use interface 144.25.142.172·          ·         Apr 3, 2008 6:30:37 PM org.collaxa.thirdparty.jgroups.protocols.UDP createSockets·          ·         INFO: socket information:·          ·         local_addr=144.25.142.172:1127, mcast_addr=228.8.15.75:45788, bind_addr=/144.25.142.172, ttl=32·         sock: bound to 144.25.142.172:1127, receive buffer size=64000, send buffer size=32000·         mcast_recv_sock: bound to 144.25.142.172:45788, send buffer size=32000, receive buffer size=64000·         mcast_send_sock: bound to 144.25.142.172:1128, send buffer size=32000, receive buffer size=64000·         Apr 3, 2008 6:30:37 PM org.collaxa.thirdparty.jgroups.protocols.TP$DiagnosticsHandler bindToInterfaces·          ·         -------------------------------------------------------·          ·         GMS: address is 144.25.142.172:1127·          ------------------------------------------------------- Example jGroups Log Entries for TCPApr 3, 2008 6:23:39 PM org.collaxa.thirdparty.jgroups.blocks.ConnectionTable start ·         INFO: server socket created on 144.25.142.172:7900·          ·         Apr 3, 2008 6:23:39 PM org.collaxa.thirdparty.jgroups.protocols.TP$DiagnosticsHandler bindToInterfaces·          ·         -------------------------------------------------------·         GMS: address is 144.25.142.172:7900------------------------------------------------------- In the log below the "socket created on" indicates that the TCP socket is established on the own node at that IP address and port the "created socket to" shows that the second node has connected to the first node, matching the logfile above with the IP address and port.Apr 3, 2008 6:25:40 PM org.collaxa.thirdparty.jgroups.blocks.ConnectionTable start ·         INFO: server socket created on 144.25.142.173:7901·          ·         Apr 3, 2008 6:25:40 PM org.collaxa.thirdparty.jgroups.protocols.TP$DiagnosticsHandler bindToInterfaces·          ·         ------------------------------------------------------·         GMS: address is 144.25.142.173:7901·         -------------------------------------------------------·         Apr 3, 2008 6:25:41 PM org.collaxa.thirdparty.jgroups.blocks.ConnectionTable getConnectionINFO: created socket to 144.25.142.172:7900  Result By reviewing the log files, you can confirm if BPEL clustering at the jGroups level is working and that the jGroup channel is communicating. Test Case 2  Test connectivity between BPEL Nodes Implementation Test connections between different cluster nodes using ping, telnet, and traceroute. The presence of firewalls and number of hops between cluster nodes can affect performance as they have a tendency to take down connections after some time or simply block them.Also reference Metalink Note 413783.1: "How to Test Whether Multicast is Enabled on the Network." Result Using the above tools you can confirm if Multicast is working  and whether BPEL nodes are commnunicating. Test Case3 Test deployment of BPEL suitcase to one BPEL node.  Implementation Deploy a HelloWorrld BPEL suitcase (or any other client specific BPEL suitcase) to only one BPEL instance using ant, or JDeveloper or via the BPEL consoleLog on to the second BPEL console to check if the BPEL suitcase has been deployed Result If jGroups has been configured and communicating correctly, BPEL clustering will allow you to deploy a suitcase to a single node, and jGroups will notify the second instance of the deployment. The second BPEL instance will go to the DB and pick up the new deployment after receiving notification. The result is that the new deployment will be "deployed" to each node, by only deploying to a single BPEL instance in the BPEL cluster. Test Case 4  Test to see if the BPEL server failsover and if all asynch processes are picked up by the secondary BPEL instance Implementation Deploy a 2 Asynch process: A ParentAsynch Process which calls a ChildAsynchProcess with a variable telling it how many times to loop or how many seconds to sleepA ChildAsynchProcess that loops or sleeps or has an onAlarmMake sure that the processes are deployed to both serversShut down one BPEL serverOn the active BPEL server call ParentAsynch a few times (use the load generation page)When you have enough ParentAsynch instances shut down this BPEL instance and start the other one. Please wait till this BPEL instance shuts down fully before starting up the second one.Log on to the BPEL console and see that the instance were picked up by the second BPEL node and completed Result The BPEL instance will failover to the secondary node and complete the flow ESB test cases This section covers the use cases involved with testing an ESB cluster. For this section please Normal 0 false false false EN-CA X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin-top:0cm; mso-para-margin-right:0cm; mso-para-margin-bottom:10.0pt; mso-para-margin-left:0cm; line-height:115%; mso-pagination:widow-orphan; font-size:11.0pt; mso-bidi-font-size:12.0pt; font-family:"Calibri","sans-serif"; mso-fareast-language:EN-US;} follow Metalink Note 470267.1 which covers the basic tests to verify your ESB cluster.

  Read the article

• Parallelism in .NET – Part 11, Divide and Conquer via Parallel.Invoke

  - by Reed

  Many algorithms are easily written to work via recursion.  For example, most data-oriented tasks where a tree of data must be processed are much more easily handled by starting at the root, and recursively “walking” the tree.  Some algorithms work this way on flat data structures, such as arrays, as well.  This is a form of divide and conquer: an algorithm design which is based around breaking up a set of work recursively, “dividing” the total work in each recursive step, and “conquering” the work when the remaining work is small enough to be solved easily. Recursive algorithms, especially ones based on a form of divide and conquer, are often a very good candidate for parallelization. This is apparent from a common sense standpoint.  Since we’re dividing up the total work in the algorithm, we have an obvious, built-in partitioning scheme.  Once partitioned, the data can be worked upon independently, so there is good, clean isolation of data. Implementing this type of algorithm is fairly simple.  The Parallel class in .NET 4 includes a method suited for this type of operation: Parallel.Invoke.  This method works by taking any number of delegates defined as an Action, and operating them all in parallel.  The method returns when every delegate has completed: Parallel.Invoke( () => { Console.WriteLine("Action 1 executing in thread {0}", Thread.CurrentThread.ManagedThreadId); }, () => { Console.WriteLine("Action 2 executing in thread {0}", Thread.CurrentThread.ManagedThreadId); }, () => { Console.WriteLine("Action 3 executing in thread {0}", Thread.CurrentThread.ManagedThreadId); } ); .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } Running this simple example demonstrates the ease of using this method.  For example, on my system, I get three separate thread IDs when running the above code.  By allowing any number of delegates to be executed directly, concurrently, the Parallel.Invoke method provides us an easy way to parallelize any algorithm based on divide and conquer.  We can divide our work in each step, and execute each task in parallel, recursively. For example, suppose we wanted to implement our own quicksort routine.  The quicksort algorithm can be designed based on divide and conquer.  In each iteration, we pick a pivot point, and use that to partition the total array.  We swap the elements around the pivot, then recursively sort the lists on each side of the pivot.  For example, let’s look at this simple, sequential implementation of quicksort: public static void QuickSort<T>(T[] array) where T : IComparable<T> { QuickSortInternal(array, 0, array.Length - 1); } private static void QuickSortInternal<T>(T[] array, int left, int right) where T : IComparable<T> { if (left >= right) { return; } SwapElements(array, left, (left + right) / 2); int last = left; for (int current = left + 1; current <= right; ++current) { if (array[current].CompareTo(array[left]) < 0) { ++last; SwapElements(array, last, current); } } SwapElements(array, left, last); QuickSortInternal(array, left, last - 1); QuickSortInternal(array, last + 1, right); } static void SwapElements<T>(T[] array, int i, int j) { T temp = array[i]; array[i] = array[j]; array[j] = temp; } Here, we implement the quicksort algorithm in a very common, divide and conquer approach.  Running this against the built-in Array.Sort routine shows that we get the exact same answers (although the framework’s sort routine is slightly faster).  On my system, for example, I can use framework’s sort to sort ten million random doubles in about 7.3s, and this implementation takes about 9.3s on average. Looking at this routine, though, there is a clear opportunity to parallelize.  At the end of QuickSortInternal, we recursively call into QuickSortInternal with each partition of the array after the pivot is chosen.  This can be rewritten to use Parallel.Invoke by simply changing it to: // Code above is unchanged... SwapElements(array, left, last); Parallel.Invoke( () => QuickSortInternal(array, left, last - 1), () => QuickSortInternal(array, last + 1, right) ); } This routine will now run in parallel.  When executing, we now see the CPU usage across all cores spike while it executes.  However, there is a significant problem here – by parallelizing this routine, we took it from an execution time of 9.3s to an execution time of approximately 14 seconds!  We’re using more resources as seen in the CPU usage, but the overall result is a dramatic slowdown in overall processing time. This occurs because parallelization adds overhead.  Each time we split this array, we spawn two new tasks to parallelize this algorithm!  This is far, far too many tasks for our cores to operate upon at a single time.  In effect, we’re “over-parallelizing” this routine.  This is a common problem when working with divide and conquer algorithms, and leads to an important observation: When parallelizing a recursive routine, take special care not to add more tasks than necessary to fully utilize your system. This can be done with a few different approaches, in this case.  Typically, the way to handle this is to stop parallelizing the routine at a certain point, and revert back to the serial approach.  Since the first few recursions will all still be parallelized, our “deeper” recursive tasks will be running in parallel, and can take full advantage of the machine.  This also dramatically reduces the overhead added by parallelizing, since we’re only adding overhead for the first few recursive calls.  There are two basic approaches we can take here.  The first approach would be to look at the total work size, and if it’s smaller than a specific threshold, revert to our serial implementation.  In this case, we could just check right-left, and if it’s under a threshold, call the methods directly instead of using Parallel.Invoke. The second approach is to track how “deep” in the “tree” we are currently at, and if we are below some number of levels, stop parallelizing.  This approach is a more general-purpose approach, since it works on routines which parse trees as well as routines working off of a single array, but may not work as well if a poor partitioning strategy is chosen or the tree is not balanced evenly. This can be written very easily.  If we pass a maxDepth parameter into our internal routine, we can restrict the amount of times we parallelize by changing the recursive call to: // Code above is unchanged... SwapElements(array, left, last); if (maxDepth < 1) { QuickSortInternal(array, left, last - 1, maxDepth); QuickSortInternal(array, last + 1, right, maxDepth); } else { --maxDepth; Parallel.Invoke( () => QuickSortInternal(array, left, last - 1, maxDepth), () => QuickSortInternal(array, last + 1, right, maxDepth)); } We no longer allow this to parallelize indefinitely – only to a specific depth, at which time we revert to a serial implementation.  By starting the routine with a maxDepth equal to Environment.ProcessorCount, we can restrict the total amount of parallel operations significantly, but still provide adequate work for each processing core. With this final change, my timings are much better.  On average, I get the following timings: Framework via Array.Sort: 7.3 seconds Serial Quicksort Implementation: 9.3 seconds Naive Parallel Implementation: 14 seconds Parallel Implementation Restricting Depth: 4.7 seconds Finally, we are now faster than the framework’s Array.Sort implementation.

  Read the article

• Backing up SQL Azure

  - by Herve Roggero

  That's it!!! After many days and nights... and an amazing set of challenges, I just released the Enzo Backup for SQL Azure BETA product (http://www.bluesyntax.net). Clearly, that was one of the most challenging projects I have done so far. Why??? Because to create a highly redundant system, expecting failures at all times for an operation that could take anywhere from a couple of minutes to a couple of hours, and still making sure that the operation completes at some point was remarkably challenging. Some routines have more error trapping that actual code... Here are a few things I had to take into account: Exponential Backoff (explained in another post) Dual dynamic determination of number of rows to backup  Dynamic reduction of batch rows used to restore the data Implementation of a flexible BULK Insert API that the tool could use Implementation of a custom Storage REST API to handle automatic retries Automatic data chunking based on blob sizes Compression of data Implementation of the Task Parallel Library at multiple levels including deserialization of Azure Table rows and backup/restore operations Full or Partial Restore operations Implementation of a Ghost class to serialize/deserialize data tables And that's just a partial list... I will explain what some of those mean in future blob posts. A lot of the complexities had to do with implementing a form of retry logic, depending on the resource and the operation.

  Read the article

• Success Quote: A Hybrid Approach for Success

  - by Lauren Clark

  We recently received this quote from a project that successfully used OUM: “On our project, we applied a combination of the Oracle Unified Method (OUM) and the client's methodology. The project was organized by OUM's phases and a subset of OUM's processes, tasks, and templates. Using a hybrid of the two methods resulted in an implementation approach that was optimized for the client-specific requirements for this project." This hybrid approach is an excellent example of using OUM in the flexible and scalable manner in which it was intended. The project team was able to scale OUM to be fit-for-purpose for their given situation. It's great to see how merging what was needed out of OUM with the client’s methodology resulted in an implementation approach that more closely aligned to the business needs. Successfully scaling OUM is dependent on the needs of the particular project and/or engagement. The key is to use no more than is necessary to satisfy the requirements of the implementation and appropriately address risks. For more information, check out the "Tailoring OUM for Your Project" page, which can be accessed by first clicking on the "OUM should be scaled to fit your implementation" link on the OUM homepage and then drilling into the link on the subsequent page. Have you used OUM in conjunction with a partner or customer methodology? Please share your experiences with us.

  Read the article

• Talking JavaOne with Rock Star Charles Nutter

  - by Janice J. Heiss

  JavaOne Rock Stars, conceived in 2005, are the top rated speakers from the JavaOne Conference. They are awarded by their peers who through conference surveys recognize them for their outstanding sessions and speaking ability. Over the years many of the world’s leading Java developers have been so recognized.We spoke with distinguished Rock Star, Charles Nutter. A JRuby Update from Charles NutterCharles Nutter of Red Hat is well known as a lead developer of JRuby, a Ruby implementation of Java that is tightly integrated with Java to allow for the embedding of the interpreter into any Java application with full two-way access between the Java and the Ruby code. Nutter is giving the following sessions at this year’s JavaOne: CON7257 – “JVM Bytecode for Dummies (and the Rest of Us Too)” CON7284 – “Implementing Ruby: The Long, Hard Road” CON7263 – “JVM JIT for Dummies” BOF6682 – “I’ve Got 99 Languages, but Java Ain’t One” CON6575 – “Polyglot for Dummies” (Both with Thomas Enebo) I asked Nutter, to give us the latest on JRuby. “JRuby seems to have hit a tipping point this past year,” he explained, “moving from ‘just another Ruby implementation’ to ‘the best Ruby implementation for X,’ where X may be performance, scaling, big data, stability, reliability, security, and a number of other features important for today's applications. We're currently wrapping up JRuby 1.7, which improves support for Ruby 1.9 APIs, solves a number of user issues and concurrency challenges, and utilizes invokedynamic to outperform all other Ruby implementations by a wide margin. JRuby just gets better and better.” When asked what he thought about the rapid growth of alternative languages for the JVM, he replied, “I'm very intrigued by efforts to bring a high-performance JavaScript runtime to the JVM. There's really no reason the JVM couldn't be the fastest platform for running JavaScript with the right implementation, and I'm excited to see that happen.”And what is Nutter working on currently? “Aside from JRuby 1.7 wrap-up,” he explained, “I'm helping the Hotspot developers investigate invokedynamic performance issues and test-driving their new invokedynamic code in Java 8. I'm also starting to explore ways to improve the general state of dynamic languages on the JVM using JRuby as a guide, and to help the JVM become a better platform for all kinds of languages.”

  Read the article

< Previous Page | 35 36 37 38 39 40 41 42 43 44 45 46  | Next Page >