I am interested in the pros/cons of refactoring a large application (industrial monitoring software) into a bunch of Libraries / NuGet Packages, rather than as stand-alone Services. The perception is that they're almost identical, i.e. a service can be built as either a Library hosted within the application, or it can be built as a Web Service and hosed externally to the application. The only difference is integration (code level vs. network SOAP or REST traffic). I'm not sure it's that straightforward, looking for pros/cons of each.
The perception is that they're almost identical...
If you squint your eyes hard enough I think they might look identical. But only up to a point.
You implement some functionality in a web service, you can implement it in a library too. The web service
provides an API to call, the library provides an API to call too. You invoke the web service, you can invoke
the library too.
But there is stuff you can't do by replacing a web service with
a library.
Your libraries are in .dll files? How is a PHP application going to use them? Or a Java application? SOAP/REST are client agnostic. Your web service can use any technology stack, your client can use any technology stack.
With libraries you are stuck to the same technology you used to write the library in.
Web services are individually deployable components. Fix bugs for the web service, deploy, and your 10 clients get the fix on the next call. Now fix some bugs for the library then update all 10 clients to use
the new versions of the libraries.
How about security? A web service can have a database for example. You don't get access to the database directly, you do so
with the API the web service provides you. You have a library that uses a database? Guess where the connection
string, username and password are found.
How about scalability? If you have a web service you can scale it horizontally and your clients immediately benefit
from the improvement. If the library is in the client how are you going to scale it? By scaling the clients?
You can find similar responses elsewhere (e.g. Web Service vs. Shared Library), but the point to take home is that it depends on what you are building and for whom. One is not a general replacement for the other, and vice-versa.
If you are building a single monolithic application that makes use of all of these components then maybe it does not matter for you if the components are microservices or hosted libs. With libs it might actually be easier to build. But if your plan is to actually provide services to other clients, then hosted libs won't work all the way.
Related
I'm building a POS(Point of sale) web application using Django for my client. This is a single instance web app.
I just want to know, should I build APIs to separate Backend and Frontend?
Whare are the benefits to build the APIs?
First and most importantly, each design decision you take for your application should be for a reason. There is no good or bad approaches, however, the approach you decide on should be based on your use-case.
I don't see any problem for having a monolithic Django application to service a single/multiple POS with limited number of users. Tens of thousands of Django applications out-there not using APIs and performing well.
That being said, if you are developing a POS solution and planning to offer it to public as SAAS (software as a service), then you might need to consider using APIs, have plans for scaling/support/patching/monitoring/upgrade/ CI and CD activities. On the other hand, if might be planning to have a Mobile application in the future; That might be another good reason to have APIs.
I am looking for the best way to add support for a REST based Web Service to an MFC C++ Application. This is legacy code that has been around a long time.
I need to have the Application provide the Web Service. I've researched this, and the articles are about consuming Web Services in a C++ App. I envision having to monitor a port and respond to the HTTP requests. It would probably make more sense to tie in with WCF and IIS, but I am not sure how to move forward.
I already have a tool of "making" this App into a Windows Service. I realize it would be better to have a "real" app, and this is what I have to work with. There is so much logic in this code and we are limited in development time, so we are taking the fastest approach possible. We also cannot use DCOM, as the Web Service will be called by a Linux based system.
Any suggested articles would be much appreciated.
For C++ you can test two libraries WSF/C and gSoap, another alternative is to develop it with C# and WCF , and communicate between C# and C++ using C++\CLI.
I use the Poco Net c++ library for this as it supports both HTTP client and server functionality. I considered various other libraries including Boost based stuff, but other than Poco I struggled to find everything I needed in one tidy package. It also has some easy to use mutex support which became essential for me as soon as I added a web server front-end to what had previously been a simple single threaded console app. Poco is cross platform but I mainly use it windows and it has served me well to date. I'm not very clued up on MFC so I don't know how it would get on in that department though.
I am looking for any information on the Microsoft TFS Web Services. First I know accessing the Microsoft TFS Web Services directly is not supported and Microsoft provides no documentation for doing this. Therefore I am not expecting any Microsoft support or assistance here.
I know all about the .Net API available for TFS which only works on Microsoft Operating Systems. I have used these many times on Windows, however I need to do non-Windows work to access TFS, I cannot use .Net and I cannot use a Proxy (or "shim") to be installed on a Windows computer to provide Web Services for the .Net API.
I know Teamprise reversed engineered the web services and they successfully used this knowledge to make a very good cross platform Team Explorer and command line implementation in Java to access TFS. So good in fact they were purchased by Microsoft and the product rebranded and rereleased as Microsoft Visual Studio Team Explorer Everywhere.
I have also tested the .Net API against Mono on several non-windows platforms and they are not compatible. The initial NTLMv2 authentication is using calls not supported by Mono. They appear to be, understandably, making Win32 specific calls for NTLMv2 support.
Therefore before I go to the trouble of reverse engineering them for myself, and dealing with NTLMv2 to do it. I am hoping that there is some hidden or buried information on the web that someone may have documented some portion of the web services for TFS from 2005, 2008 and/or 2010.
Please no comments or posts about how this is not recommended or supported by Microsoft, that I should find a way to use the .Net API, or suggesting the Proxy/Shim is the best solution. I am fully aware of the Microsoft's official stance on this, and what the supported workarounds would be.
I'm not aware of any documentation for the TFS web services, but I can share some tips on calling them.
The NTLM authentication you mention is really a separate layer: you must authenticate to IIS before it lets you call TFS web services. I'm not aware of any Open Source software that will do NTLM auth for you, but TFS 2010 makes it easy to enable "Negotiate" authentication (SPNEGO on Wikipedia, Authentication by using Kerberos Ticket on MSDN). Negotiate supports both NTLM and Kerberos subsystems, and there may be some existing software you can use to drive it using the system's Kerberos libraries (I think curl does it). If you had to build it yourself, it would probably be easier to go the Negotiate-with-Kerberos route.
Once you're authenticated, you can start calling services. Start by pulling down the WSDL for each service (stick a "?wsdl" suffix on each endpoint URI). Hop over to where TFS is installed and explore the web application directory for endpoints. There are several versions of some endpoints for back compat with TFS 2005 and 2008, but usually new versions are not redundant (they add new stuff). You might have a favorite SOAP client library already (there are many for Java), but I can't really recommend any because we wrote our own at Teamprise.
Services like version control, build, and common structure are easy to discover via WSDL. Most the operations have obvious names, but the complex type fields are often super-abbreviated. The best way to figure which methods to call when is to watch the VS TFS client or TEE with Fiddler or Wireshark or some other HTTP inspection program. TFS VC does do things like file uploads/downloads outside the web services (watch a network trace to see the multi-part MIME upload process and be sure you're sending the right values if you implement this).
A note of caution on the work item tracking web service: this one is going to be extremely hard to master. The WIT design involves the client pre-querying the server for large amounts of schema-less metadata, which is saved on the client (but refreshed incrementally as more web service calls are made). This metadata drives all the client side behavior about work items (what fields are in a work item type, the type of a field, which values are allowed in fields, the rules that run when they change, etc.) and it will take a long time and serious study to build the client behavior to bring a work item to life. Once you have a work item, sending it to the server for update via web services is easy.
It's a lot of work, but it's possible to do incrementally, for example, if you only need some VC features. The TEE team is working on making access from other platforms easier. Please contact Martin Woodward (martin.woodward#microsoft.com) if you have any questions or suggestions in this area.
There is a Java version of the TFS SDK that will run on Linux, Mac, and Windows. It is the SDK that Teamprise uses.
http://blogs.msdn.com/b/bharry/archive/2011/05/16/announcing-a-java-sdk-for-tfs.aspx
Coding directly against the TFS webservices is not supported (even though people have done it). MSFT could break the interface without letting you know in a service pack or other hotfix. Sometimes there aren't other options, but if the Java SDK works for you, I'd try to use that first.
There is good documentation now: https://www.visualstudio.com/integrate/get-started/rest/basics
OK, this is impossible, but I will try to explain the situation here.
Let's say we have cases, that we need a fast setup of a web server in order to have a simple soap web service running (querying a db and so on).
In VS though, upon debugging a web project, it creates a quick ASP.NET development server without relying on the actuall IIS that might be installed on the PC.
Is there any project that does something like that?
This would be ideal for small projects, where a simple executable would get a server ready to go and would allow web services to be executed right away.
I have looked at some stuff over the net like http://msdn.microsoft.com/en-us/magazine/cc163879.aspx and http://mtaulty.com/CommunityServer/blogs/mike_taultys_blog/archive/2004/05/24/4479.aspx but they seem to be quite outdated and i am not sure how well they work (havent tested them thoroughly)
EDIT: I have build an application like the one you suggest. However, how can i implement HTTP GET/POST requests to the service using this method? I tried using WebGet after my operationcontract but it didnt work. For example, my service is at http://localhost:8080/Service and i would like to use it such as http://localhost:8080/Service/getMethod?x=2.
I believe that the development server used by Visual Studio is based off of the Cassini code base (of which there is a fork here). I also found this article on hosting the asp.net runtime. It was also created a while back (2004), but has been updated since (2008). I think a lot of the core concepts are probably still the same.
Another approach would be a roll-your-own web server using the HttpListener class. This could take some work if you want to use it for hosting asmx type services, but if you were doing RESTful services, it isn't too bad of an option (this is actually how RavenDB works if you are not hosting it under IIS).
A WCF service can be hosted in almost any kind of application, including a Windows Service or a console application. There is no need for a web server at all.
Alright,
i've done it so im posting it here to help anyone who has issues with similar problems.
Create your WCF Service file as usual and then by using ServiceHost (or WebServiceHost) you can easily create a WCF service.
In order to use GET http requests to make it simple to communicate with mobile devices (such as iphone) you can use WebGet above your service methods and make sure you manually add a service endpoint using WebHttpBinding for WebServiceHost or WebHttpBinding with an WebHttpBehavior for ServiceHost.
Then you can call your service methods such as http://localhost:port/webhttpendpointaddress/mymethod?x=2.
I want to create a simple web service which I shall then make available to third parties to host. The third parties will want to host the service as part of their own systems be that Windows, Linux or Solaris. They will want the web service to be made available to them as a pre compiled binary which they can run without dependencies on further third party software other than the OS itself.
So I am looking for a library, preferably in C / C++, which I can use to create a stand alone web service. I don't want to buy, install configure and manage some Application Server middleware such as Web Sphere, nor do my customers. I am happy to have a RESTFUl service rather than using SOAP.
I'm sure there are more than options available each with their own pro's and con's. A comparison of the options would be fantastic.
Regards
You may want to take a look at Webtoolkit (or 'wt'): www.webtoolkit.eu