Chapter 1. Introduction to REST
For those of us with computers, the World Wide Web is an intricate part of our lives. We use it to read the newspaper in the morning, pay our bills, perform stock trades, and buy goods and services, all through the browser, all over the network. “Googling” has become a part of our daily vocabulary as we use search engines to do research for school, find what time a movie is playing, or to just search for information on old friends. Door-to-door encyclopedia salesmen have gone the way of the dinosaur as Wikipedia has become the summarized source of human knowledge. People even socialize over the network using sites like Facebook and MySpace. Professional social networks are sprouting up in all industries as doctors, lawyers and all sorts of professionals use them to collaborate. As programmers, the Web is an intricate part of our daily jobs. We search for and download open source libraries to help us develop applications and frameworks for our companies. We build web-enabled applications so that anybody on the Internet or intranet can use a browser to interact with our systems.
Really, most of us take the Web for granted. Have you, as a programmer, sat down and tried to understand why the Web has been so successful? How has it grown from a simple network of researchers and academics to an interconnected worldwide community? What properties of the Web make it so viral?
One man, Roy Fielding, did ask these questions in his doctoral thesis, “Architectural Styles and the Design of Network-based Software Architectures.” In it, he identifies specific architectural principles that answer the following questions:
Why is the Web so prevalent and ubiquitous?
What makes the Web scale?
How can I apply the architecture of the Web to my own applications?
- Addressable resources
The key abstraction of information and data in REST is a resource, and each resource must be addressable via a URI (Uniform Resource Identifier).
- A uniform, constrained interface
Use a small set of well-defined methods to manipulate your resources.
- Communicate statelessly
Stateless applications are easier to scale.
- Hypermedia As The Engine Of Application State (HATEOAS)
Let your data formats drive state transitions in your applications.
For a PhD thesis, Roy’s paper is actually very readable and, thankfully, not very long. It, along with Leonard Richardson and Sam Ruby’s book RESTful Web Services (O’Reilly), is an excellent reference for understanding REST. I will give a much briefer introduction to REST and the Internet protocol it uses (HTTP) within this chapter.
REST and the Rebirth of HTTP
REST isn’t protocol-specific, but when people talk about REST, they usually mean REST over HTTP. Learning about REST was as much of a rediscovery and reappreciation of the HTTP protocol for me as learning a new style of distributed application development. Browser-based web applications see only a tiny fraction of the features of HTTP. Non-RESTful technologies like SOAP and WS-* use HTTP strictly as a transport protocol and thus use a very small subset of its capabilities. Many would say that SOAP and WS-* use HTTP solely to tunnel through firewalls. HTTP is actually a very rich application protocol that provides a multitude of interesting and useful capabilities for application developers. You will need a good understanding of HTTP in order to write RESTful web services.
HTTP is a synchronous request/response-based application network protocol used for distributed, collaborative, document-based systems. It is the primary protocol used on the Web, in particular, by browsers such as Firefox, MS Internet Explorer, Safari, and Netscape. The protocol is very simple: the client sends a request message made up of the HTTP method being invoked, the location of the resource you are interested in invoking, a variable set of headers, and an optional message body that can basically be anything you want, including HTML, plain text, XML, JSON, even binary data. Here’s an example:
GET /resteasy/index.html HTTP/1.1 Host: jboss.org User-Agent: Mozilla/5.0 Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 Accept-Language: en-us,en;q=0.5 Accept-Encoding: gzip,deflate
Your browser would send this request if you wanted to look at http://jboss.org/resteasy/index.html. GET is the method we
are invoking on the server.
/resteasy/index.html is the object we are
interested in. HTTP/1.1 is the version of the protocol.
Accept-Encoding are all message headers. There
is no request body, as we are querying information from the server.
The response message from the server is very similar. It contains the version of HTTP we are using, a response code, a short message that explains the response code, a variable set of optional headers, and an optional message body. Here’s the message the server might respond with using the previous GET query:
HTTP/1.1 200 OK X-Powered-By: Servlet 2.4; JBoss-4.2.2.GA Content-Type: text/html <head> <title>JBoss RESTEasy Project</title> </head> <body> <h1>JBoss RESTEasy</h1> <p>JBoss RESTEasy is an open source implementation of the JAX-RS specification...
The response code of this message is 200, and the status message is “OK.” This code means that the request was processed successfully and that the client is receiving the information it requested. HTTP has a large set of response codes. They can be informational codes like 200, “OK,” or error codes like 500, “Internal Server Error.” Visit www.w3.org/Protocols/rfc2616/rfc2616-sec10.html for a more complete and verbose listing of these codes.
This response message also has a message body that is a chunk of
HTML. We know it is HTML by the
RESTful Architectural Principles
Roy Fielding’s PhD thesis describing REST was really an explanation of why the human readable Web had become so pervasive in the past 18 years. As time went on though, programmers started to realize that they could use the concepts of REST to build distributed services and model service-oriented architectures (SOAs).
The idea of SOA is that application developers design their systems as a set of reusable, decoupled, distributed services. Since these services are published on the network, conceptually, it should be easier to compose larger and more complex systems. SOA has been around for a long time. Developers have used technologies like DCE, CORBA, and Java RMI to build them in the past. Nowadays, though, when you think of SOA, you think of SOAP-based web services.
While REST has many similarities to the more traditional ways of writing SOA applications, it is very different in many important ways. You would think that a background in distributed computing would be an asset to understanding this new way of creating web services, but unfortunately this is not always the case. The reason is that some of the concepts of REST are hard to swallow, especially if you have written successful SOAP or CORBA applications. If your career has a foundation in one of these older technologies, there’s a bit of emotional baggage you will have to overcome. For me, it took a few months of reading, researching, and intense arguing with REST evangelists (aka RESTafarians). For you, it may be easier. Others will never pick REST over something like SOAP and WS-*.
Let’s examine each of the architectural principles of REST in detail and why they are important when writing a web service.
Addressability is the idea that every object and resource in your system is reachable through a unique identifier. This seems like a no-brainer, but if you think about it, standardized object identity isn’t available in many environments. If you have tried to implement a portable J2EE application, you probably know what I mean. In J2EE, distributed and even local references to services are not standardized, which makes portability really difficult. This isn’t such a big deal for one application, but with the new popularity of SOA, we’re heading to a world where disparate applications must integrate and interact. Not having something as simple as standardized service addressability adds a whole complex dimension to integration efforts.
In the REST world, addressability is managed through the use of URIs. When you make a request for information in your browser, you are typing in a URI. Each HTTP request must contain the URI of the object you are requesting information from or posting information to. The format of a URI is standardized as follows:
scheme is the protocol you
are using to communicate with. For RESTful web services, it is usually
host is a DNS name or IP address. It is
followed by an optional
is numeric. The
port represent the location of your resource
on the network. Following
port is a
path expression. This
path expression is a set of text segments
delimited by the “/” character. Think of the
path expression as a directory list of a file
on your machine. Following the path expression is an optional query
string. The “?” character separates the path from the query string. The
query string is a list of parameters represented as name/value pairs.
Each pair is delimited with the “&” character. Here’s an example
query string within a URI:
A specific parameter name can be repeated in the query string. In this case, there are multiple values for the same parameter.
The last part of the URI is the
fragment. It is delimited by a “#” character.
The fragment is usually used to point to a certain place in the document
you are querying.
Not all characters are allowed within a URI string. Some characters must be encoded using the following rules. The characters a–z, A–Z, 0–9, ., -, *, and _ remain the same. The space character is converted to +. The other characters are first converted into a sequence of bytes using specific encoding scheme. Next, a two-digit hexadecimal number prefixed by % represents each byte.
Using a unique URI to identify each of your services makes each of your resources linkable. Service references can be embedded in documents or even email messages. For instance, consider the situation where somebody calls your company’s help desk with a problem related to your SOA application. A link could represent the exact problem the user is having. Customer support can email the link to a developer who can fix the problem. The developer can reproduce the problem by clicking on the link. Furthermore, the data that services publish can also be composed into larger data streams fairly easily:
<order id="111"> <customer>
http://customers.myintranet.com/customers/32133</customer> <order-entries> <order-entry> <quantity>5</quantity> <product>
In this example, an XML document describes an e-commerce order entry. We can reference data provided by different divisions in a company. From this reference, we can not only obtain information about the linked customer and products that were bought, but we also have the identifier of the service this data comes from. We know exactly where we can further interact and manipulate this data if we so desired.
The Uniform, Constrained Interface
The REST principle of a constrained interface is perhaps the hardest pill for an experienced CORBA or SOAP developer to swallow. The idea behind it is that you stick to the finite set of operations of the application protocol you’re distributing your services upon. This means that you don’t have an “action” parameter in your URI and use only the methods of HTTP for your web services. HTTP has a small, fixed set of operational methods. Each method has a specific purpose and meaning. Let’s review them:
GET is a read-only operation. It is used to query the server for specific information. It is both an idempotent and safe operation. Idempotent means that no matter how many times you apply the operation, the result is always the same. The act of reading an HTML document shouldn’t change the document. Safe means that invoking a GET does not change the state of the server at all. This means that, other than request load, the operation will not affect the server.
PUT requests that the server store the message body sent with the request under the location provided in the HTTP message. It is usually modeled as an insert or update. It is also idempotent. When using PUT, the client knows the identity of the resource it is creating or updating. It is idempotent because sending the same PUT message more than once has no effect on the underlying service. An analogy is an MS Word document that you are editing. No matter how many times you click the Save button, the file that stores your document will logically be the same document.
POST is the only nonidempotent and unsafe operation of HTTP. Each POST method is allowed to modify the service in a unique way. You may or may not send information with the request. You may or may not receive information from the response.
OPTIONS is used to request information about the communication options of the resource you are interested in. It allows the client to determine the capabilities of a server and a resource without triggering any resource action or retrieval.
There are other HTTP methods (like TRACE and CONNECT), but they are unimportant when designing and implementing RESTful web services.
You may be scratching your head and thinking, “How is it possible to write a distributed service with only four to six methods?” Well…SQL only has four operations: SELECT, INSERT, UPDATE, and DELETE. JMS and other Message Oriented Middleware (MOM) really only have two logical operations: send and receive. How powerful are these tools? For both SQL and JMS, the complexity of the interaction is confined purely to the data model. The addressability and operations are well defined and finite and the hard stuff is delegated to the data model (in the case of SQL) or the message body (in the case of JMS).
Why Is the Uniform Interface Important?
If you have a URI that points to a service, you know exactly which methods are available on that resource. You don’t need an IDL-like file describing which methods are available. You don’t need stubs. All you need is an HTTP client library. If you have a document that is composed of links to data provided by many different services, you already know which method to call to pull in data from those links.
HTTP is a very ubiquitous protocol. Most programming languages have an HTTP client library available to them. So, if your web service is exposed over HTTP, there is a very high probability that people who want to use your service will be able to do so without any additional requirements beyond being able to exchange the data formats the service is expecting. With CORBA or WS-*, you have to install vendor-specific client libraries as well as loads and loads of IDL or WSDL-generated stub code. How many of you have had a problem getting CORBA or WS-* vendors to interoperate? It has traditionally been very problematic. The WS-* set of specifications has also been a moving target over the years. So with WS-* and CORBA, you not only have to worry about vendor interoperability, you also have to make sure that your client and server are using the same specification version of the protocol. With REST over HTTP, you don’t have to worry about either of these things and can just focus on understanding the data format of the service. I like to think that you are focusing on what is really important: application interoperability, rather than vendor interoperability.
Because REST constrains you to a well-defined set of methods, you have predictable behavior that can have incredible performance benefits. GET is the strongest example. When surfing the Internet, have you noticed that the second time you browse to a specific page it comes up faster? This is because your browser caches already visited pages and images. HTTP has a fairly rich and configurable protocol for defining caching semantics. Because GET is a read method that is both idempotent and safe, browsers and HTTP proxies can cache responses to servers, and this can save a huge amount of network traffic and hits to your website. Add HTTP caching semantics to your web services and you have an incredibly rich way of defining caching policies for your services. We will discuss HTTP caching in detail within Chapter 10.
It doesn’t end with caching, though. Consider both PUT and DELETE. Because they are idempotent, neither the client nor the server has to worry about handling duplicate message delivery. This saves a lot of bookkeeping and complex code.
The third architectural principle of REST is that your services should be representation-oriented. Each service is addressable through a specific URI and representations are exchanged between the client and service. With a GET operation, you are receiving a representation of the current state of that resource. A PUT or POST passes a representation of the resource to the server so that the underlying resource’s state can change.
In a RESTful system, the complexity of the client-server interaction is within the representations being passed back and forth. These representations could be XML, JSON, YAML, or really any format you can come up with.
With HTTP, the representation is the message body of your request
or response. An HTTP message body may be in any format the server and
client want to exchange. HTTP uses the
Content-Type header to tell the client or
server what data format it is receiving. The
Content-Type header value string is in the
Multipurpose Internet Mail Extension (MIME) format. The MIME format is
type is the main format family
subtype is a category.
Optionally, the MIME type can have a set of name/value pair properties
delimited by the “;” character. Some examples are:
text/plain text/html application/xml text/html; charset=iso-8859-1
One of the more interesting features of HTTP that leverages MIME
types is the capability of the client and server to negotiate the
message formats being exchanged between them. While not used very much
by your browser, HTTP content negotiation is a very powerful tool when
writing web services. With the
header, a client can list its preferred response formats. Ajax clients
can ask for JSON, Java for XML, Ruby for YAML. Another thing this is
very useful for is versioning of services. The same service can be
available through the same URI with the same methods (GET, POST, etc.),
and all that changes is the MIME type. For example, the MIME type could
application/vnd+xml for an old
service while newer services could exchange
application/vnd+xml;version=1.1 MIME types.
You can read more about all these concepts in Chapter 8.
All in all, because REST and HTTP have a layered approach to addressability, method choice, and data format, you have a much more decoupled protocol that allows your service to interact with a wide variety of different clients in a consistent way.
The fourth RESTful principle I will discuss is the idea of statelessness. When I talk about statelessness, though, I don’t mean that your applications can’t have state. In REST, stateless means that there is no client session data stored on the server. The server only records and manages the state of the resources it exposes. If there needs to be session-specific data, it should be held and maintained by the client and transferred to the server with each request as needed. A service layer that does not have to maintain client sessions is a lot easier to scale, as it has to do a lot fewer expensive replications in a clustered environment. It’s a lot easier to scale up, because all you have to do is add machines.
A world without server-maintained session data isn’t so hard to imagine if you look back 12–15 years ago. Back then, many distributed applications had a fat GUI client written in Visual Basic, Power Builder, or Visual C++ talking RPCs to a middle tier that sat in front of a database. The server was stateless and just processed data. The fat client held all session state. The problem with this architecture was an IT operations one. It was very hard for operations to upgrade, patch, and maintain client GUIs in large environments. Web applications solved this problem because the applications could be delivered from a central server and rendered by the browser. We started maintaining client sessions on the server because of the limitations of the browser. Now, circa 2008, with the growing popularity of Ajax, Flex, and Java FX, the browsers are sophisticated enough to maintain their own session state like their fat-client counterparts in the mid-’90s used to do. We can now go back to that stateless scalable middle tier that we enjoyed in the past. It’s funny how things go full circle sometimes.
The final principle of REST is the idea of using Hypermedia As The Engine Of Application State (HATEOAS). Hypermedia is a document-centric approach with the added support for embedding links to other services and information within that document format. I did indirectly talk about HATEOAS in Addressability when I discussed the idea of using hyperlinks within the data format received from a service.
One of the uses of hypermedia and hyperlinks is composing complex sets of information from disparate sources. The information could be within a company intranet or dispersed across the Internet. Hyperlinks allow us to reference and aggregate additional data without bloating our responses. The e-commerce order in Addressability is an example of this:
<order id="111"> <customer>
http://customers.myintranet.com/customers/32133</customer> <order-entries> <order-entry> <quantity>5</quantity> <product>
In that example, links embedded within the document allowed us to bring in additional information as needed. Aggregation isn’t the full concept of HATEOAS, though. The more interesting part of HATEOAS is the “engine.”
The engine of application state
If you’re on Amazon.com buying a book, you follow a series of links and fill out one or two forms before your credit card is charged. You transition through the ordering process by examining and interacting with the responses returned by each link you follow and each form you submit. The server guides you through the order process by embedding where you should go next within the HTML data it provides your browser.
This is very different from the way traditional distributed applications work. Older applications usually have a list of precanned services they know exist and interact with a central directory server to locate these services on the network. HATEOAS is a bit different because with each request returned from a server it tells you what new interactions you can do next, as well as where to go to transition the state of your applications.
For example, let’s say we wanted to get a list of products available on a web store. We do an HTTP GET on http://example.com/webstore/products and receive back:
<products> <product id="123"> <name>headphones</name> <price>$16.99</price> </product> <product id="124"> <name>USB Cable</name> <price>$5.99</price> </product> ... </products>
This could be problematic if we had thousands of products to send back to our client. We might overload it, or the client might wait forever for the response to finish downloading. We could instead list only the first five products and provide a link to get the next set:
<products> <link rel="next" href="http://example.com/webstore/products?startIndex=5"/> <product id="123"> <name>headphones</name> <price>$16.99</price> </product> ... </products>
When first querying for a list of products, clients don’t have
to know they’re only getting back a list of five products. The data
format can tell them that they didn’t get a full set and to get the
next set follow a specific link.
next link could get
them back a new document with additional links:
<products> <link rel="previous" href="http://example.com/webstore/products?startIndex=1"/> <link rel="next" href="http://example.com/webstore/products?startIndex=5"/> <product id="128"> <name>stuff</name> <price>$16.99</price> </product> ... </products>
In this case, there is the additional state transition of
previous so that clients can browse
an earlier part of the product list. The
previous links seem a bit trivial, but
imagine if we had other transition types like
This sort of approach gives the server a lot of flexibility, as it can change where and how state transitions happen on the fly. It could provide new and interesting opportunities to surf to. In Chapter 9, we’ll dive into HATEOAS again.
REST identifies the key architectural principles of why the Web is so prevalent and scalable. The next step in the evolution of the Web is to apply these principles to the semantic Web and the world of web services. REST offers a simple, interoperable, and flexible way of writing web services that can be very different than the RPC mechanisms like CORBA and WS-* that so many of us have had training in. In the next chapter we will apply the concepts of REST by defining a distributed RESTful interface to an existing business object model.