Sastry,
My personal feeling is that a HTTP REST interface is such a large feature that this might have to wait for the next
version of JMS, or perhaps be taken forward in a separate JSR. However I don't want to close down discussion on the
subject and would welcome the EG having an opportunity to discuss this issue (you'll have seen my separate email asking
EG members to list which possible new features they would like to discuss - I presume this would be *your* priority).
Although I am asking for a "proposal", I'm thinking of something fairly short and high-level (and perhaps with a few
examples) that will help us understand what type of API we're talking about and what it might be used for. I'm not
asking for a fully-worded out specification.
Thanks,
Nigel
On 30/03/2012 18:47, Sastry Malladi wrote:
> I agree that we need to have a concrete proposal and I don't mind volunteering to work on such a proposal - The only
> catch is I'm going on a sabbatical for 1 month starting next week and if it still helps, I can put a proposal together
> after I get back, if timelines match.
>
> You are correct that someone can just implement the HTTP REST interface only and re-use some existing JMS provider and
> in fact, I'm currently implementing such a provider. So from a compliance perspective, implementing either interface
> or both should be allowed.
>
> So, if you think there is a possibility that we will add this to the current spec, then I can work on a proposal when
> I get back. Let me know.
>
> thanks,
> Sastry
> ------------------------------------------------------------------------------------------------------------------------
> *From:* Nigel Deakin <nigel.deakin_at_oracle.com>
> *To:* jsr343-experts_at_jms-spec.java.net
> *Sent:* Wednesday, March 28, 2012 7:58 AM
> *Subject:* [jms-spec users] [jsr343-experts] Re: (JMS_SPEC-7) Provide HTTP Binding
>
> (Sorry for the delay)
>
> Yes, this all sounds very interesting. I think what this needs is for one or more individuals to take this forward to
> the next level of detail - perhaps listing the various JMS functions it would support and what kind of HTTP
> action/request/callback this would correspond to. Even if this doesn't make it into JMS 2.0 (and we are all still
> deciding priorities - see separate email) it would be really great to have one or more proposals on the table.
>
> Effectively we'd be defining the API for a web application which could use the normal JMS (Java) API to communicate
> with the actual JMS provider (at least for the reference implementation). So we would probably want to allow people to
> implement it and become spec-compliant without the need to implement the JMS provider as well (and vice versa). We
> might decide that this is best achieved as a separate JSR but I'm happy to treat it within the scope of JMS for now.
>
> Whether or not a Java EE application server would be required to support the HTTP interface could be considered
> separately. That would be something to discuss with the Java EE expert group at a later stage.
>
> You mentioned "local calls to the embedded JMS server". I know that some JMS providers do allow the JMS "server" to be
> embedded in the same JVM as the application, and provide a JMS API which uses method calls rather than the network to
> communicate with it. However not all do and this is not required by the spec. So we'd probably need to leave it to the
> individual HTTP protocol implementation to decide whether to take advantage of such a feature for a particular JMS
> provider.
>
> Nigel
>
>
> On 06/03/2012 05:40, Sastry Malladi wrote:
>> Nigel,
>> Yes, I was thinking about defining a REST/HTTP interface for the subset of the API - Transactionality and ordering of
>> messages etc. is not important to expose via this interface. I was thinking of exposing basic Restful resources : A
>> Queue, A topic, sending and receiving messages to them, specifying a subscription (a filter condition) and
>> preferences for message delivery - pull Vs push (using callback endpoints, websockets or otherwise). In addition a
>> management and monitoring interface as well. Specifying the QoS (best effort/real-time, at-least-once and
>> at-most-once semantics).
>>
>> What I mean by "embeddable mode" friendly is that the JMS server should have a embeddable mode option, in which the
>> Restful resources implementation can make local java calls to the embedded JMS server without going through the
>> network. There already is a network hop through HTTP and we want to avoid a second hop.
>>
>> Sastry
>>
>> ------------------------------------------------------------------------------------------------------------------------
>> *From:* Nigel Deakin <nigel.deakin_at_oracle.com> <mailto:nigel.deakin_at_oracle.com>
>> *To:* jsr343-experts_at_jms-spec.java.net <mailto:jsr343-experts_at_jms-spec.java.net>
>> *Sent:* Monday, March 5, 2012 3:17 AM
>> *Subject:* [jsr343-experts] Re: (JMS_SPEC-7) Provide HTTP Binding
>>
>> Sastry,
>>
>> Thanks for your comments.
>>
>> As you say, stateful JMS concepts such as connections and sessions are difficult to express in a HTTP/REST interface.
>> The same applies to the JMS concepts of transactions and client acknowledgement.
>>
>> Were you thinking of a HTTP/REST interface which mapped onto a subset of existing JMS functionality (e.g. no message
>> order guarantees, no transactions) or do yo think we need to define some new concepts in JMS which map more easily to
>> a internet environment, such as new QoS options?
>>
>> You also mention "embedded mode". Can you clarify what you mean by this, and what kind of changes would be needed to
>> make JMS "embeddable mode" friendly?
>>
>> Thanks,
>>
>> Nigel
>>
>>
>> On 05/03/2012 07:16, Sastry Malladi wrote:
>>> Hi,
>>> I'm sorry for being silent so far on the mailing list. I've recently joined the expert group (in Jan) - I'm am an
>>> Architect at eBay driving the eBay middleware platform, including
>>> a next gen messaging platform.
>>>
>>> In the internet world, the messaging systems tend to serve a more "disconnected world" - meaning that there is no
>>> concept of a "connection" or "session" to the broker - Applications just publish and receive messages (millions of
>>> them) at will, with different QoS and delivery channels. These applications are also typically polyglot (written in
>>> multiple different languages) and almost always use HTTP/REST to communicate, although the likes of SPDY and
>>> Websockets are also partly in use and upcoming.
>>>
>>> Although this JSR is for Java API for messaging, I'd still strongly request you to consider including an HTTP, more
>>> importantly a simple REST interface to the messaging API, for wider adoption. We don't have to make implementing
>>> the REST interface mandatory for compliance, but I bet most providers will implement, as the world is moving towards
>>> that direction. This is what we are doing as well. The REST service implementation may choose to use the underlying
>>> Java API in embedded mode, so we need to make sure the Java API is "embeddable mode" friendly.
>>>
>>> thanks,
>>> Sastry
>>>
>>> ------------------------------------------------------------------------------------------------------------------------
>>> *From:* Nigel Deakin <nigel.deakin_at_oracle.com> <mailto:nigel.deakin_at_oracle.com>
>>> *To:* jsr343-experts_at_jms-spec.java.net <mailto:jsr343-experts_at_jms-spec.java.net>
>>> *Sent:* Friday, March 2, 2012 8:44 AM
>>> *Subject:* [jsr343-experts] (JMS_SPEC-7) Provide HTTP Binding
>>>
>>> I refer to this JIRA issue logged by fribeiro in June 2011.
>>>
>>> Provide HTTP Binding
>>> http://java.net/jira/browse/JMS_SPEC-7
>>>
>>> "If none is available from another organization, I think the JCP should provide (maybe in a separate JSR) a standard
>>> HTTP binding for JMS, given how often these technologies are used together."
>>>
>>> A discussion took place in the comments on that issue between the proposer and various EG members. I'd like to bring
>>> that discussion to the EG properly so we can formally decide how to handle this.
>>>
>>> I think that this proposal is is essentially proposing that JMS defines some kind of HTTP binding (protocol, really)
>>> to a JMS provider. I can well imagine that this is a common requirement: I know at least two JMS providers that
>>> provide a HTTP protocol and I'm sure there are others.
>>>
>>> I think there are a number of issues here:
>>>
>>> 1. Whether a standard HTTP protocol to JMS is required
>>> 2. If so, whether it belongs in JMS or in some other specification
>>> 3. Whether there is scope to enhance the existing JMS (Java) API to make it easier to deliver a HTTP binding
>>>
>>> Defining a standard HTTP protocol sounds, on the face of it, a good idea. It would be necessary to decide what JMS
>>> features could be made available using HTTP - some, like message order or transactions, would probably rely on the
>>> concept of there being some kind of client state maintained between requests.
>>>
>>> Then there's the question of whether this is should be defined as part of JMS, as a separate JCP specification, or
>>> under the auspices of some other body.
>>>
>>> As a general rule the JCP is "for developing standard technical specifications for Java technology", but defining a
>>> HTTP protocol for JMS is certainly not out of the question. Some would probably recommend that it be defined at
>>> OASIS, or even IETF. But if it needs to align strongly with the Java API then that might be enough justification to
>>> develop it as part of JMS.
>>>
>>> We would need to consider what the compatibility requirements would be for the HTTP protocol? Would all JMS
>>> products be required to include a REST server that supported the protocol?
>>>
>>> We would also need to consider where the HTTP protocol would sit the JMS architecture. Would the JMS server support
>>> the HTTP protocol or would we be defining a separate server that accepted HTTP requests and translated them to the
>>> native network protocol for the JMS server, perhaps by just translating them into JMS API calls?
>>>
>>> My feeling is that we would never want to make it mandatory for a JMS provider to directly support the HTTP
>>> protocol, and that it should be possible to implement it as a separate component interfacing with the JMS provider
>>> using the standard JMS API (if it requires proprietary API then it isn't really a JMS binding). This suggests to me
>>> that this belongs in a separate JSR.
>>>
>>> I'm also mindful that this would be a significant piece of work and there's not going to be time to deliver in the
>>> JMS 2.0 timescales in any case.
>>>
>>> So my proposal is that we take the decision to not attempt to define a HTTP protocol for JMS 2.0. We can leave the
>>> issue open, but it is likely that a HTTP protocol would need to be delivered as a separate JSR.
>>>
>>> Discussion welcome.
>>>
>>> Now let's consider the point (3) above, whether there is scope to enhance the existing JMS (Java) API to make it
>>> easier to deliver a HTTP binding. I think the answer to this is definitely yes, and we welcome proposals. (I think
>>> we already have one in JMS_SPEC-5, which we'll discuss separately, and there may be others).
>>>
>>> Nigel
>>>
>>>
>>
>>
>
>