I'm designing a new API and I'm struggling with some decisions. I've read tons of blogs on SOAP vs REST and I used the popular APIs (Paypal, Amazon, etc.) as my guidelines.
I ended up with 2 endpoints in my API: one for SOAP and one for REST (XML). The SOAP one looks pretty good, but the XML interface looks somewhat strange. I'm calling it "strange" because I ended up with namespaces in some of my tags. For example:
[sample1]
<EnvelopeRequest xmlns:c1='http://foobar/CarrierX'>
<Weight>1.0</Weight>
<PostmarkDate>5/3/2013</PostmarkDate>
<c1:ShippingMethod>Ground</c1:ShippingMethod>
<c1:Notification>a#b.com</c1:Notification>
</EnvelopeRequest>
[sample2]
<EnvelopeRequest xmlns:cs='http://foobar/SpecialCarrier'>
<Weight>1.0</Weight>
<PostmarkDate>5/3/2013</PostmarkDate>
<cs:Shape>Flat</cs:Shape>
</EnvelopeRequest>
The reason the XML interface has namespaces is because it is auto-generated from the class definition (which has some inheritance). We are using WCF btw. That works just fine for SOAP (the WSDL is derived from the same class), because SOAP hides all the ugliness in the client proxies. However, after looking at many REST/XML services, I don't think I've seen namespaces being used too often. This also kinda scares me because I'm thinking that I would love to have a JSON interface in the near future, and JSON doesn't support namespaces.
My decision to make the API SOAP friendly came from the fact that many of our customers use Enterprise solutions which thrive on SOAP. But lately, with the growing popularity of Python and Ruby, which new clients seem to adopt more often, I'm starting to second guess my initial decision. The main thing that bothers me is the namespaces in the XML interface, but is it really an issue? Are namespaces in a REST/XML API such a big no-no that I should change my design?
If I do change my design, then my (2 previous) requests would look like so:
[sample1]
<EnvelopeRequest>
<Weight>1.0</Weight>
<PostmarkDate>5/3/2013</PostmarkDate>
<CarrierX>
<ShippingMethod>Ground</ShippingMethod>
<Notification>a#b.com</Notification>
</CarrierX>
</EnvelopeRequest>
[sample2]
<EnvelopeRequest>
<Weight>1.0</Weight>
<PostmarkDate>5/3/2013</PostmarkDate>
<SpecialCarrier>
<Shape>Flat</Shape>
</SpecialCarrier>
</EnvelopeRequest>
And yes, this would allow me to have a JSON interface in the future.
Removing namespaces would be a problem if by doing so you create the possibility of ambiguity in a given message. Is it possible for someone somewhere to create an EnvelopeRequest message with a Shape element that might be interpreted (by code or by people reading the message) in more than one way? The reason to introduce namespaces is to preclude this possibility. Tools like WCF's auto-generator are not able to answer this question in the general case so they err on the side of caution.
Only you can know the set of possible valid messages. In my experience, it's usually preferable to remove namespaces for the sake of not confusing your users/clients. There are a few reasons why I might change that preference:
I expect my message format to be used widely and intermixed with other formats. (A good example is the Atom syndication format)
I'm using someone else's widely used (and namespaced) format and planning to intermix it with my own (e.g. embedding XHTML inside my message).
I expect to embed a message of a given format inside a message of the same format (e.g. XSLT stylesheets that generate XSLT stylesheets).
In that latter case, you might find it convenient (though not absolutely necessary) to use namespaces to separate the inner message from the message that is carrying it by using different prefixes. I don't think any of these cases apply very often.
I would ponder why you have namespace in the first place, those are some strange payloads.
But, disregarding that, no, the namespaces are not a big deal. Namespaces almost inevitably run afoul with XPath and XSL (since they tend to be namespace aware), but when consuming the document wholesale, a lot of times folks just ignore the namespace component completely, so in the end there's no difference.
I would clean up the namespaces for the sake of cleaning them up semantically, but not necessarily for the sake of the consumers. From a practical stand point, it's not that big a deal.
we are currently thinking about how to design an interface for other systems.
My co-worker would like to implement a generic interface (for e.g. doIt(JSONArray)) where you put the desired information you would like to do inside a JSONObject, so that calls would e.g. look like this:
doIt('{"method":"getInformation", "id":"1234", "detailLevel": "2"}')
doIt('{"method":"getEmployeeInfo", "EmployeeId":"4567", "company": "Acme Inc."}')
(i used ' and " in this example just for demonstration purposes. I know that i had to escape the " in the real system).
This method will then be accessable via http, so that i would like http://mysite/doIt?parm={JSONObject}
My approach is to use different interfaces with their respective parameters so that I would have a getInformation(1234,2) and a getEmployeeInfo(4567,"Acme Inc.") interface. So for access via http my scheme would look like: http://mysite/getInformation?id=1234&detailLevel=2 and http://mysite/getEmployeeInfo?employeeId=4567&company=acmeinc.
For the clients accessing our service we want to provide special libraries that encapsulate the bevahiour. E.g. there will a client java-lib which translates a client-call getEmployeeInfo(..) either to
http://mysite/doIt?parm={'{"method":"getEmployeeInfo", "EmployeeId":"4567", "company": "Acme Inc."}'}
or to
http://mysite/getEmployeeInfo?employeeId=4567&company=acmeinc.
and then return the result.
So for clients it will be completely transparent how the backend works if they use the library which handles the "translation".
What do you think are the pros and cons of each idea? I like my approach better because it looks "cleaner". But that is just a feeling which is difficult to argue about. Perhaps you can give me (or him) some thoughts about the design and also touch areas (scalability, security,...) or provide useful links about this matter
I'd probably vote for the JSON solution, even if they are more or less equivalent. (Both easily extendable, standard, future-proof solutions.)
The reasons for choosing JSON:
There are a plethora of different libraries for different platforms that help you build correct objects, check that the string data is valid, etc.
Unmarshalling of JSON data into objects. Some libraries (for example Gson) can automatically marshal and unmarshal JSON into objects. Saves you from writing your own code, and you get the benefit of using code that has been tested by others.
Support for new interfaces. Suppose that you change your transport method to sockets, ftp(!) or whatever. You could still send the JSON objects to you backend using another transport.
I realize this question is old, but I think the answers here would guide developers down the wrong path.
In my experience you should always lean towards the more specific methods. Generic methods are difficult to test, difficult to wrap your head around and provide no (or minimal) IDE/compiler support. Such an api you are describing does not tell the user anything about what it will do.
Your own suggested api design is much better.
That being said, its a balancing act.
The JSON solution could be better because you can send complex object easier
But here it's just a little syntax detail, let the boss choose (or do a vote) and build your software.
Related:
Why would one use REST instead of Web services?
When deciding whether to implement a web service using SOAP or REST (by which I mean HTTP/XML in a RESTful manner) what should I be aware of and what should I be thinking of? I presume that this isn't a one size fits all thing so how do I choose which to use.
The two protocols have very different uses in the real world.
SOAP(using WSDL) is a heavy-weight XML standard that is centered around document passing. The advantage with this is that your requests and responses can be very well structured, and can even use a DTD. The downside is it is XML, and is very verbose. However, this is good if two parties need to have a strict contract(say for inter-bank communication). SOAP also lets you layer things like WS-Security on your documents. SOAP is generally transport-agnostic, meaning you don't necessarily need to use HTTP.
REST is very lightweight, and relies upon the HTTP standard to do it's work. It is great to get a useful web service up and running quickly. If you don't need a strict
API definition, this is the way to go. Most web services fall into this category. You can version your API so that updates to the API do not break it for people using old versions(as long as they specify a version). REST essentially requires HTTP, and is format-agnostic(meaning you can use XML, JSON, HTML, whatever).
Generally I use REST, because I don't need fancy WS-* features. SOAP is good though if you want computers to understand your webservice using a WSDL. REST specifications are generally human-readable only.
The following links provide useful information about WSDL vs REST including Pros and Cons
A couple of key points are that
1) SOAP was designed for a distributed computing environment where as REST was designed for a point to point environment.
2) WADL can be used to define the interface for REST services.
http://www.ajaxonomy.com/2008/xml/web-services-part-1-soap-vs-rest
http://ajaxonomy.com/2008/xml/web-services-part-2-wsdl-and-wadl
Regarding WSDL (meaning "SOAP") as being "heavy-weight". Heavy matters how? If the toolset is doing all the "heavy lifting" for you, then why does it matter?
I have never yet needed to consume a complicated REST API. When I do, I expect I'll wish for a WSDL, which my tools will gladly convert into a set of proxy classes, so I can just call what appear to be methods. Instead, I suspect that in order to consume a non-trivial REST-based API, it will be necessary to write by hand a substantial amount of "light-weight" code.
Even when that's all done, you still will have translated human-readable documentation into code, with all the attendant risk that the humans read it wrong. Since WSDL is a machine-readable description of the service, it's much harder to "read it wrong".
Just a note: since this post, I have had the opportunity to work with a moderately complicated REST service. I did, indeed, wish for a WSDL or the equivalent, and I did, indeed, have to write a lot of code by hand. In fact, a substantial part of the development time was spent removing the code duplication of all the code that called different service operations "by hand".
This probably really belongs as comments in several of the above posts, but I don't yet have the rep to do that, so here goes.
I think it is interesting that a lot of the pros and cons often cited for SOAP and REST have (IMO) very little to do with the actual values or limits of the two technologies. Probably the most cited pro for REST is that it is "light-weight" or tends to be more "human readable". At one level this is certainly true, REST does have a lower barrier to entry - there is less required structure than SOAP (though I agree with those who have said that good tooling is largely the answer here - too bad much of the SOAP tooling is pretty dreadful).
Beyond that initial entry cost however, I think the REST impression comes from a combination of the form of the request URLs and the complexity of the data exchanged by most REST services. REST tends to encourage simpler, more human readable request URLs and the data tends to be more digestable as well. To what extent however are these inherent to REST and to what extent are they merely accidental. The simpler URL structure is a direct result of the architecture - but it could be equally well applied to SOAP based services. The more digestable data is more likely to be a result of the lack of any defined structure. This means you'd better keep your data formats simple or you are going to be in for a lot of work. So here SOAP's additional structure, which should be a benefit is actually enabling sloppy design and that sloppy design then gets used as a dig against the technology.
So for use in the exchange of structured data between computer systems I'm not sure that REST is inherently better than SOAP (or visa-versa), they are just different. I think the comparison above of REST vs SOAP to dynamic vs. static typing is a good one. Where dyanmic languages tend to run in to trouble is in long term maintenance and upkeep of a system (and by long term I'm not talking a year or 2, I'm talking 5 or 10). It will be interesting to see if REST runs into the same challenges over time. I tend to think it will so if I were building a distributed, information processing system I would gravitate to SOAP as the communication mechanism (also because of the tranmission and application protocol layering and flexibility that it affords as has been mentioned above).
In other places though REST seems more appropriate. AJAX between the client and its server (regardless of payload) is one major example. I don't have much care for the longevity of this type of connection and ease of use and flexibility are at a premimum. Similarly if I needed quick access to some external service and I didn't think I was going to care about the maintainability of the interaction over time (again I'm assuming this is where REST is going to end up costing me more, one way or another), then I might choose REST just so I could get in and out quickly.
Anyway, they are both viable technologies and depending on what tradeoffs you want to make for a given application they can serve you well (or poorly).
REST is not a protocol; It's an architectural style. Or a paradigm if you want. That means that it's a lot looser defined that SOAP is. For basic CRUD, you can lean on standard protocols such as Atompub, but for most services you'll have more commands than just that.
As a consumer, SOAP can be a blessing or a curse, depending on the language support. Since SOAP is very much modelled on a strictly typed system, it works best with statically typed languages. For a dynamic language it can easily become crufty and superfluous. In addition, the client-library support isn't that good outside the world of Java and .NET
To me we should be careful when we use the word web service. We should all the time specify if we are speaking of SOAP web service, REST web service or other kind of web services because we are speaking about different things here and people don't understand anymore if we named all of them web services.
Basically SOAP web services are very well established for years and they follow a strict specification that describe how to communicate with them based on the SOAP specification.
Now REST web services are a bit newer and basically looks like simpler because they are not using any communication protocol. Basically what you send and receive when you use a REST web service is plain XML. People like it because they can parse the xml the way they want without having to deal with a more sophisticated communication protocol like SOAP.
To me REST services are almost like if you would create a servlet instead of a SOAP web service. The servlet get data in and return data out. The format of the data are xml based. We can also imagine to use something else than xml if we want. For instance tags could be used instead of xml and that would be not REST anymore but something else (Could be even lighter in term of weight because xml is not light by nature). Would we call that still a web service? Yes we could but that will not follow any current standard and this is the main issue here if we start to call everything web services but we can do it the way we want then we are loosing on the interoperability side of the things. That means that the format of the data that is exchanged with the web service is not standardized anymore. That requires then that server and client agree on the format of the data whereas with SOAP this is all predefined already and server and client can interoperate without to know each other because they follow the same standard.
What people don't like with SOAP is that they have hard time to understand it and they cannot generate the queries manually. Computers can do that very well however so this is where we need to be clear: are web services queries and response supposed to be used directly by the end users or do we agree that web services are underneath API called by computer systems based on some normalized standards?
SOAP: It can be transported via SMTP also, means we can invoke the service using Email simple text format also
It needs additional framework/engine should be in web service consumer machine to convert SOAP message to respective objects structure in various languages.
REST: Now WSDL2.0 supports to describe REST web service also
We can use when you want to make your service as lightweight, example calling from mobile devices like cell phone, pda etc...
for enterprise systems in which your system is confined within your corporations, its easier and proper to use soap because you are almost in control of clients. it's easier since there a variety of tools which creates classes (proxies) and looks like you are doing your regular OOP which matches your java or .net environment (in which most corporates use).
I would use REST for internet facing applications for exposing interfaces (like twitter api) since clients can be using javascripts or html or others in which typing is not strict. REST being more liberal makes more sense.
Also for internet facing clients (world wide web), its easier to parse json or xml coming out of a rest interface rather than a purely xml coming from a soap interface. it's hard to use proxies on javascript and javascript does not naturally support objects. If you are using REST with javascript, you would just usually parse the json string and you're off. internet facing interfaces are usually very simple (so most of the time its simple parsing) and does not usually demand consistency that is why REST is adequate enough.
For enterprise applications I don't think REST is adequate because transactions, security, strict typing, schemas play a very important in enterprise applications development that is why SOAP is more suited for them.
My conclusion is that SOAP is for Enterprise systems, REST is for the Internet or WWW.
You can use it interchangeably but you may find yourself having a difficult time eventually not using the correct tool for the job.
sorry for my bad english.
In defence of REST it closely follows the principles of HTTP and addressability e.g. read operations use GET, update operations use POST etc. I find this to be a far cleaner approach. The Oreilly book RESTful Web Services explains this far better than I can, if you read it I think you would prefer the REST approach
The toolset on the client side would be one. And the familiarity with SOAP services the other. More and more services are going the RESTful route these days, and testing such services can be done with simple cURL examples.
Although, it's not all that difficult to implement both methods and allow for the widest utilization from clients.
If you need to pick one, I'd suggest REST, it's easier.
The previous answers contain a lot of information, but I think there is a philosophical difference that hasn't been pointed out. SOAP was the answer to "how to we create a modern, object-oriented, platform and protocol independent successor to RPC?". REST developed from the question, "how to we take the insights that made HTTP so successful for the web, and use them for distributed computing?"
SOAP is a about giving you tools to make distributed programming look like ... programming. REST tries to impose a style to simplify distributed interfaces, so that distributed resources can refer to each other like distributed html pages can refer to each other. One way it does that is attempt to (mostly) restrict operations to "CRUD" on resources (create, read, update, delete).
REST is still young -- although it is oriented towards "human readable" services, it doesn't rule out introspection services, etc. or automatic creation of proxies. However, these have not been standardized (as I write). SOAP gives you these things, but (IMHO) gives you "only" these things, whereas the style imposed by REST is already encouraging the spread of web services because of its simplicity. I would myself encourage newbie service providers to choose REST unless there are specific SOAP-provided features they need to use.
In my opinion, then, if you are implementing a "greenfield" API, and don't know that much about possible clients, I would choose REST as the style it encourages tends to help make interfaces comprehensible, and easy to develop to. If you know a lot about client and server, and there are specific SOAP tools that will make life easy for both, then I wouldn't be religious about REST, though.
You can easily transition your WSDL-spewing WCF web components to other uses just by changing your configuration settings. You can go across HTTP and then also named pipes, tcp, custom protocols, etc without having to change your code. I believe WCF components may also be easier to set up for stuff like security, two-way calling, transactions, concurrency, etc.
REST pretty much limits you to HTTP (which is fine in many cases).
I know that this discussion is an old one, but after reading all the answers and commented, I believe that everyone missed the most important point about the difference between the 2 systems: SOAP uses complex types to not only give you the data, but validate it and keep it in the strict type designation it was defined for. A WSDL tells you what the data format is, what the data type is, allows you to add reg-ex pattern-style rules, and defines how many times a piece of data must be, and may be, allowed in a request/response.
Rest on the other-hand has none of these mechanisms.
SOAP is complex and heavy because it allows you to send complex heavy hierarchical data. REST is plain text, with the origin and endpoint sorting out the rules.
SOAP is business independent, because it has all the data rules embedded in the document.
The difference between SOAP and REST is that SOAP is a self-contained business oriented schema. REST is a text document.
I have a website (ASP.NET) and some winforms(.Net 2.0) for a project (written in C#). I use the webservice (IIS6) for task that both require like sending email inside the business.
I think Webservice is nice but I would like from your experience what should and what should not be in a webservice?
In My Opinion:
Web services should be reserved for code that
You either can't or don't want to distribute; or,
code that needs to seriously scale up.
One example is custom business logic that multiple applications need access to.
Code you don't want to put into web services include:
code that is performance based;
code that applies only to the application in question.
Well it sounds like you have a limited Service Oriented Architecture (at least, that's what I think you're getting at), which according to Gartner means you'll be rich soon. :)
I find that the benefit of SOA for me really comes down to the heterogeneity of the systems involved (sounds like yours doesn't qualify there because it's all .NET), and the negative of SOA is primarily because of the verbose nature of XML. True, you don't need XML for SOA, but it's the current majority, IMHO.
But if you're not concerned about the bandwidth/parsing penalties, who cares? Maybe you're not piping through 10,000 service calls a minute. With this style of implementation, you're following DRY, just with a WS instead of a sub, and you're adhering to a standard that is by nature compatible with multiple systems.
There's worse approaches.
It seems like the new trend for web services/SOA is to more or less expose a light-weight middle tier that your host application can use. Instead of having individual method calls exposed through a service (as in your example), SOA-oriented applications have extensive Data/Operation contracts that act as the "traditional" middle tier assembly.
As little as possible, while still being useful.
By default, DON'T put every field of the return objects in the return data, and DON'T expose every method of an existing class.
read this too...
What is the minimum set of HTTP verbs that a server should allow for a web service to be classed as RESTful?
What if my hoster doesn't permit PUT and DELETE?
Is this actually important, can I live happily ever after with just GET and POST ?
Update: Thanks for the answers folks, Roger's answer was probably best because of the link to the Bill Venners and Elliotte Rusty Harold interview. I now get it.
Yes, you can live without PUT and DELETE.
This article tells you why:
http://www.artima.com/lejava/articles/why_put_and_delete.html
While to true RESTafrians this may be heresy, in the real world you do what you can, with what you have. Be as rational as you can and as consistent with your own convention as you can, but you can definitely build a good RESTful system without P and D.
rp
You can also use X-Http-Verb-Override:DELETE inst. of HTTP DELETE. This is also usefull for Silverlight clients who cant change the HTTP verbs and only support GET and POST...
If you just use GET and POST, it's still RESTful. Your web service may only do things which only required GET or POST, so that's fine.
REST allows for breaking protocol convention if the implementations of the protocol are broken (so that the only non-standard things you do are to get around the broken parts of the implementation). So it is allowable within REST to use some other method to represent generally unsupported verbs like DELETE or PUT.
edit: Here is a quote from Fielding, who is the one that created and defined REST:
A REST API should not contain any changes to the communication protocols aside from filling-out or fixing the details of underspecified bits of standard protocols, such as HTTP’s PATCH method or Link header field. Workarounds for broken implementations (such as those browsers stupid enough to believe that HTML defines HTTP’s method set) should be defined separately, or at least in appendices, with an expectation that the workaround will eventually be obsolete. [Failure here implies that the resource interfaces are object-specific, not generic.]
Today's web browsers only handle GETS + POSTS. In Rails, for example, PUTS + DELETES are "faked" through hidden form fields.
Unless your framework has some workaround to "support" PUTS + DELETES, don't worry about them for now.