Bloated Domain Objects And CQRS
 

Problem of Bloated Domain objects

In business software applications, the domain objects (entities) are used to represent the business domain. As the application grows and adds more business logic, the service layer, mappers and other patterns gets applied. This sector has held promise for many software developing companies, and has often been touted as the future of work. Often this leads to domain object becomes bloated and the related components become huge & un-maintainable.

CQRS solves the common problem of having a bloated Domain objects. The domain objects get bloated largely because of bounded context. The series of contexts which makes developers think that a single domain object is sufficient to handle all the related things. For example, a large Invoice object for handling Invoice, Shipment and handling change of address for customer . But in reality, these contexts (invoicing, shipment and change) need not be related to same Invoice entity.

What is Command, Query Responsibility segregation (CQRS)?

In order to simplify the Domain objects, CQRS proposes to have two types of domain entities.

  • those serving the command (ordering/assertion services)  – For example, SaveCustomer, CreateInvoice, ShipProduct etc
  • those serving a Query (request) – examples include GetCustomerAddress, SearchCustomer etc

With this separation, the complexity (number of fields, methods) of entities used becomes simplified. And hence the Data mapper layers & the service layers becomes more simplified.

Where can I use CQRS?

  • Largely complex system: Applying CQRS on a simple CRUD operation based system is a over kill. When there is a domain heavy system, like banking and financing systems, LOB applications where business logic, lots of boundary conditions are heavy. Where it makes DDD (Domain driven design) provides high value.
  • Situations where you will apply Microservices, Eventual consistency and Event Sourcing. When we have separation of concerns using CQRS, the microservices becomes much simpler to design and maintain. With Event sourcing we are focused on getting the data (query) from other related sources and is what CQRS propagates.

Final words

CQRS is a carefully thought out pattern for simplifying & solving large and complex systems.

50 Tips and Tricks for Web Performance
 

I came across the below video on the web performance improvement tips and tricks. This is fantastic.

In this video, Jatinder talks about Six fundamental principles for improving web application performance. He also talks a lot about how we go about Decreasing CPU time and increasing parallelism on the client machine to achieve faster web performance.

While I went through the video, I captured all the tricks he talks about. And thought will be useful for others while they watch it. Please find below the tricks.

Quickly respond to network request

  • Avoid 3XX Redirections (63% of top websites use redirect)
  • Avoid Meta refresh
  • Minimise Server time for Requests
  • Use Content distribution Networks (CDN)
  • Maximise concurrent connections.
  • Reuse connections – don’t sent connection close.
  • Know your other servers – you are only fast as your weakest link
  • Understand your network timing

Minimise bytes downloaded

  • GZIP Compression Network traffic
  • Persist application resources locally (Windows 8 applications)
  • Cache dynamic resources in the application Cache (html5 app cache)
  • Provide cacheable content
  • Send Conditional request
  • Cache Data requests (jQuery AJAX calls)
  • Standardize File name capitalization convention.

Efficiently structure markup

  • Load pages in latest browser mode
  • Use http header to specify legacy IE mode
  • Link CSS in the top of the page header, never on the bottom.
  • Avoid using @import of hierarchical styles
  • Avoid embedded and inline CSS
  • Only include necessary styles in the page.
  • Always link JS at the end of the page.
  • Avoid linking JS in the header (use the defer & async attribute)
  • Avoid Inline JS
  • Remove Duplicate code (52% of the web have duplicate code
  • Standardise on a single framework

Optimise your media usage

  • Avoid death by too many images
  • If possible use Image Sprites
  • Use png image file format
  • Use Native image resolutions
  • Replace Images with CSS3 Gradients, border Radius
  • Leverage CSS3 Transforms
  • Use Data URI’s for Small Single view images
  • Avoid complex SVG paths
  • Video : Use preview images
  • Minimize media plugin usage
  • Proactively download Future media

Write Fast JavaScript

  • Stick to Integer math (Math.Floor)
  • Minify your JavaScript
  • Initialize JavaScript on Demand
  • Minimize your DOM Interactions
  • Built in DOM methods always more efficient (firstchild, nextsibling methods are faster)
  • Use Selectors for Collection access (document.querySelectorAll)
  • Use .innerHTML to construct your page
  • Batch your markup changes
  • Maintain smaller DOM (less than 1000 elements)
  • JSON always faster than XML (possible myth)
  • Use Native JSON methods
  • Use regular expressions Sparingly

Know what your app is doing

  • Understand JavaScript Timers (setTimeout, SetInterval)
  • Combine Application Timers
  • Ensure dormant timers are not running
  • Align timers to display Frame (16.7)
  • Use window.requestAminationFrame(renderLoop) for Animations
  • Know when your application is visible (document.hidden, Visibilitychange (event))

Conclusion

The web optimisation is not easy and needs exhaustive, deep look and hopefully this check list helps while optimising your pages. Enjoy coding high performing applications. If you have more tips please provide them in the comments.

 

Micro-Services, Eventual Consistency and Event sourcing patterns
 

Microservices is a really becoming a famous architectural pattern that most of the new software written these days, just apply them. One of the most important differentiation between the traditional web services and micro-services pattern is the amount of common stuff across different subject areas. This calls for a discussion on how Eventual Consistency pattern is mandatory for successfully implementing microservices.

The micro frontend if gaining lots of popularity.  You can read about microservices principles and micro frontends at

Microservices and Microfrontends

Micro-Service Pattern

Generally, in a micro-service pattern, the API’s are split into small subject areas. For example for a CRM application, the subject areas are

  • Customer information – like name, address, email, phone
  • Appointment information – which customer, salesperson, when, where
  • Relationship management – sales/manager, what products, interests
  • Campaign data – offers, deals etc

Then micro-services are built for each of the subject areas.  The microservices are logically and physically separated from each other. ie there is no sharing (code, database, component etc) between any of these micro-services of these subject areas. Pictorially its looks something like this.

Applying Eventual Consistency Pattern

In Micro-services, there is no data that is shared across the micro services. In order to synchronize the data across these isolated storages of these services, we need to apply the eventual consistency pattern. You can read more about applying the pattern correctly here. The simpler way we can achieve consistency across these micro-services is through Event Sourcing pattern.

Event Sourcing

Event sourcing is a process of capturing application state changes in the form of events.  An example of events are customer created, customer updated, Deal created, etc.  Other systems listen to these events and they take relevant action. You can read more about event sourcing here.

Conclusion

Event sourcing is the new way of storing changes to systems and help in making micro-services eventually consistent. These patterns together form well maintainable, reliable and scalable systems in the modern world.

 

 

MEAN stack Part 2 – Yeoman and generator-angular-fullstack
 

This is the second post in the Getting started with MEAN series where we start looking at the Node ecosystem for stable web application frameworks for our application. A brief glance through the Express framework showed that express by itself would be time consuming start off with from scratch. On the other hand, the Express.js website links to a number of frameworks built on top of Express. This looked like a good place to start exploring.

Going through the list of frameworks mentioned there, however, we quickly noticed that a number of options were either targeted only (or primarily) towards APIs, lacked sufficient documentation for our needs, or did not appear to have a sufficiently large community behind them. Mind you, the assessment wasn’t entirely objective, as I did have previous experience with generator angular fullstack.

Why generator-angular-fullstack

Generator angular fullstack is a Yeoman generator that provides an application skeleton as well as sub generators that allow the developer to quickly scaffold various UI and API components. The generator supports a number of customizations through questions asked before generating the app. The ones I found most useful were:

Client Side

  • Html templating : HTML, PUG (previously Jade)
  • Angular Router : ngRouter, ui-router
  • Stylesheets : css, stylus, sass, less
  • CSS frameworks : bootstrap (with the option to include UI Bootstrap)

Server Side:

  • Scripts : javascript, Typescript
  • Databases : none, Mongodb, SQL
  • Authentication boilerplate
  • oAuth integration with Facebook, Twitter and Google Plus
  • Socket IO integration

Phew! That was a lot of options built into the generator itself, stuff that teams and products usually like to customize according to their convenience. This by itself is a strong reason to select the framework, because none of the other frameworks that we evaluated came anywhere close in terms of customizability and out of the box functionality. The generator also comes with built in hooks to deploy your application to Heroku or Openshift, although I found that part to be a little broken (more on that later)

A lot of things have changed though, since my last experience. For one, the guys over at generator-angular-fullstack added support for Typescript. Typescript in fact, is a superset of javascript, but brings a number of enhancements (including transpile time type safety) to the table. On the other hand, the generator still works with Angular 1.x, although the alpha version does support Angular 2. But then, working with the alpha version for a project needing quick turnaround times didn’t sound like an exciting thought.

Anyway, coming back to getting started with the application.

Getting Started

Installing Yeoman

 npm install -g yo

Installing the angular fullstack generator and its prerequisites

npm install -g yo gulp-cli generator-angular-fullstack

On a side note, please read node-gyp’s installation guide before going any further to get the application to run successfully. Node-gyp is required to compile native drivers for each platform and on windows, there are a couple of different ways to support the compilation process. Personally, I was able to get option 2 on the link mentioned above working (see: “Option 2: Install tools and configuration manually”)

Initializing the App

Finally! Coming down to the crux of the matter. Initializing the app

Run : yo angular-fullstack

The installer asks a number of questions about customization, including the ones mentioned above, and once the installer has completed, you have a fully functional app (with authentication if you selected it during install) ready for you to work on.

Running the app is as simple as running : gulp serve

You just need to make sure that if you selected a database, it is running on your local box so the node API can connect to it. If the DB is not found, the application will simply crash, but you already knew that would happen 🙂

Stay tuned for the next article about the getting familiar with the code generated by the Yeoman generator

Most frequently asked .NET interview questions for 5 years or less
 
  1. What is the difference between System.Array.CloneTo() and System.Array.CopyTo()?
  2. What is the difference between an interface and an abstract class? Give examples of their use in real life?
  3. Can you use “this” with the static class?
  4. What is the primary difference between “read-only” and “constants”?
  5. What is a garbage collector? Can we force the garbage collector to run? Is that guaranteed that it would run?
  6. What is public, static, void, and main?
  7. Can we execute multiple catch blocks written for the same try block?
  8. What is the difference between Finalize() and Dispose() method?
  9. What is a sealed class in C#? Explain with examples?
  10. What is the difference between a HashTable and a dictionary? When to use what? Give an example of each?
  11. What is “out” and “ref” parameter in C#?
  12. What is “protected internal” in C#?
  13. What is a “finally” in exception handling? How does it differ from finalize?
  14. What is generics?
  15. What is the difference between a connection pooling and an object pooling? How and where do you set the connection pooling parameter?  What is the default value?
  16. What is multiple inheritance? Is it allowed in C#? How do you achieve that?
  17. What is boxing and unboxing? Give examples of each?
  18. What is a copy constructor?
  19. What is a static constructor?
  20. What is the difference between overriding and overloading? Give examples?
  21. What do you mean by virtual method?
  22. What types of comments are used in C#?
  23. What is abstraction, inheritance, polymorphism, and encapsulation?
  24. What is the relation between data hiding and encapsulation?
  25. What all the access modifiers you have in C#?
Getting Started with the MEAN stack (Mongodb Express Angular Node) – What is MEAN
 

I’ve been helping out some friends with setting up a new site for their endeavors. The basic requirements were

  1. The site needs to be a ‘single page application’ (SPA)
  2. Should be light weight
  3. Should be easy to deploy to one of the many available hosted environments

MEAN stands for the set of technologies used in application development, much like the LAMP (Linux, Apache, MySql and PHP) stack before it. The technologies and frameworks involved in the MEAN stack include Mongodb, Express Js, Angular Js, and Node Js. From a comparative standpoint, I believe that MEAN is popular today, for many of the same reasons that the LAMP stack gained popularity earlier.

Despite their limitations, both development stacks received a strong push from the community, have ubiquitous ‘Hello, World!’ tutorials, are supported on a number of cheap (and sometimes free!) PAAS hosting providers, and required little to no prior experience to get off the ground, giving instant results with minimal effort. They might not be the best dev stacks of their times, but the sheer inertia from their communities has also played a major role in the two being among the most popular web development platforms in the world today.

Why use the MEAN stack

One of the primary advantages of the MEAN stack that the code is written in Javascript, through all layers of the stack. This means that the programmer’s experience with one language (Javascript), is reused for everything, including persistence and server side code (in addition to the usual client side manipulation using a UI framework, which in this case, happens to be AngularJS)

The other big advantage of having javascript at all levels in your dev stack is that objects don’t need to be marshaled and unmarshaled when moving from one layer to another, usually removing the boilerplate code responsible for such conversions. This also helps developers take advantage of Javascript’s loosely typed nature to manipulate data as and when needed. Considering that Mongodb is a document database, and ‘documents’ don’t really *need* to have a fixed schema, this lines up pretty well with Javascript’s the loosely typed nature allowing us to do some really interesting things.

One of the drawbacks that I have personally experienced, however, of the MEAN stack is that developers with prior experience primarily in relational databases tend to continue to think in terms of tables. Relating collections to tables, and documents to records isn’t wrong, conceptually, but the patterns for relational patterns don’t necessarily lend themselves favourably to NoSql databases. Relational databases tend to have long, mature histories with a standardized set of interfaces and features that a developer can reasonably expect to get if he chooses any of the competing options out there.

Sadly, in my experience, NoSql is nothing like that. Different implementations (read databses) have widely differing ideologies on the ‘Right Way’™ of doing things, and they are correct! This is mostly because each implementation seems to be built for a slightly different use case, that needs different trade-offs to make the system function efficiently (for the chosen scenario). This doesn’t mean the databases can’t be used for anything outside their original intended mainline use cases, just that the default settings for the system may not be the best for you. The minor differences in behaviors also means that again, patterns identified for one NoSql database may not be the best for a different one.

While this may change in future as the stack matures and ‘enterprise patterns’ (I hate that term, but it seems apt for this situation) emerge and become ‘common knowledge’ ingrained in a majority of the developers out there, this is not the case today.

Check out the next post on how we got started.

MEAN stack Part 2 – Yeoman and generator-angular-fullstack

Azure Functions in plain English
 

AWS Lambda and Azure functions are truly next generational in the sense of the cloud computing. These services simplify the amount of time it takes to move to a cloud-based solution. Following are some of the advantages of running services directly in functions/Lambda.

  • No infrastructure code
  • Easy scaling
  • Easy deployment
  • Gives focus on the business

In this article, we will explore how to create a simple azure function and execute it in simple 5 steps.

1.  Search for Functions in the Azure Portal and select it.

2. Click on “Create” and Create functions form should appear.

3. Enter the App Name, Subscription, Location and select “Create”. After it is created, you should see a page like this.

4. Select “Functions” from the left menu and on the right side select “Webhook + API” and click on “Create this function”.

5. Below screen appears where you can write code for the function and deploy the same. Edit code and select “Run” to update the deployment and run.

That’s all and functions are as simple as that. Hope this article helps in creating functions with Azure simpler. Please feel free to post your questions in the comments section.

Simple introduction to NServiceBus
 

What is NServiceBus

NServiceBus is a Microsoft .Net based framework for implementing the Enterprise Service Bus pattern. Its a framework for implementing reliable messaging through various transports without having to write boilerplate code. NServiceBus also provides facilities for pub/sub, retries, format abstraction, encryption, workflow orchestration, and scalability.  Below are a few highlights of the system.

  • Seamlessly work with various messaging systems (Transport)
  • Automatic Serialization & De-Serialisation
  • Enforcing a Bus architecture
  • Transaction Scope (Unit of Work)
  • A simple model to handle long-running transactions (Sagas)

Building a simple Unicast bus using NServiceBus

Let’s use .Net core 2.0 for building a simple bus. Everything in NServicebus can be configured using a simple class called “EndpointConfiguration”. An endpoint is a service that uses NServicebus framework for handling messages.

var endpointConfiguration = new EndpointConfiguration("UnicastBus");

NServiceBus uses persistence to store all the configuration and temporary data so that the data is not lost while there is a failure (fallacies of Network). We can configure NServiceBus persistence using the below code. We are using an In-memory persistence for the sample app. But in real applications, other reliable persistence like SQL, Raven, Azure Service fabric etc. will be used.

endpointConfiguration.UsePersistence<InMemoryPersistence>();

NserviceBus is a framework and is not a message queues on its own. However, it uses an Adapter pattern to fit any open source & commercial queuing systems. In our case, we are wiring it up with RabbitMQ (an open-source message queuing system written in Erlang).

var tranportExtensions = endpointConfiguration.UseTransport<RabbitMQTransport>();

tranportExtensions.ConnectionString("host=localhost;username=guest;password=guest");

RabbitMQ also has a concept of routing topology where a user can configure how data is passed between exchanges & queues. In our case, we are going to use ‘direct routing’ topology.

tranportExtensions.UseDirectRoutingTopology();

var routing = tranportExtensions.Routing();
routing.RouteToEndpoint(assembly: typeof(Order).Assembly,destination: "Sales");

Through this configuration, we are pushing all messages coming to “Unicast Bus” exchange to “Sales” exchange. That’s it, we are good to start the Unicast bus.

var endpointInstance = await Endpoint.Start(endpointConfiguration).ConfigureAwait(false);

Now, sending messages using the bus that we just created is simple.

endpointInstance..Send(“My First Message using NServiceBus”);

To see the result, we should run the program and look for messages in the ‘Sales’ queue in RabbitMQ management tool.

Conclusion:

NServiceBus is one of the frameworks for doing reliable messaging for handling the fallacies of distributed computing. This is just a gentle introduction to NServiceBus. We shall see more advanced concepts in future posts.