On Jan 22, 2014, at 1:26 PM, Emmanuel Bernard <emmanuel(a)hibernate.org> wrote:
Conceptually I like the grouping API better than AtomicMap as I don’t
have to rely on a specific Infinispan type.
We do use FineGrainedAtomicMap both for the entity and the association persistence (not
AtomicMap).
So you don't use the AtomicMap(vs FGAM) at all? Is there any place in which you
require a lock in the whole map to be acquired?
It is particularly critical for how we store the association
navigation information. I don’t want one update to literally prevent the whole association
from being updated. This is the same semantic a RDBMS has and that’s why Manik and I
designed the FGAM requirements.
So my question is what are the differences between the grouping API and the FGAM in
particular for:
- the amount of data sent back and forth (seems like grouping is sending the data
naturally per key as “delta compared to the group"
- we'll only send the (group, key, value) for every group write. The same amount of
info is sent for an FGAM.put
- the locking level when a new entry is added to the FGAM / Grouping
API
- the locking level when a new entry is removed to the FGAM / Grouping API
- the locking level when a new entry is updated to the FGAM / Grouping API
- in the case of grouping the lock object is the tuple (group, key), so the lock
granularity is the same as FGAM, which under the hood build an synthetic lock object based
on (FGAM, innerKey)
- the overall network verbosity
- same
- does grouping offer the same repeatable read protection that
AtomicMap offers within a transaction?
- yes, and it actually has a clearer semantics, as the grouping API is entirely built on
top of the basic cache operations, instead of being a first class citizen with it's
own transaction semantics.
I think retrying as a transaction workaround is quite fragile. We can offer it as a
solution but supporting or encouraging it is another story. Unless each OGM nodes do
behave like a transaction but that would be wrong. I am also concerned about reading data
form a group that are inconsistent.
The grouping API I'm suggesting offers an nicer alternative to the (FG)AM approach
that would also work over HotRod. ATM there's no TX support for HotRod so it seems
like to be able to support the HotRod and OGM integration fully, we'd need HR
transactions as well. Do you think the integration can be done in steps:
1. add grouping over hotrod and integrate with OGM
2. add tx and make the integration use it
Or we should wait till 2. and then proceed with the integration?
Emmanuel
On 21 Jan 2014, at 16:07, Mircea Markus <mmarkus(a)redhat.com> wrote:
> Hi Emmanuel,
>
> Just had a good chat with Davide on this and one solution to overcome the shortcoming
you mentioned in the above email would be to enhance the hotrod client to support
grouping:
>
> RemoteClient.put(G g, K k, V v); //first param is the group
> RemoteClinet.getGroup(G g) : Map<K,V>;
>
> It requires an enhancement on our local grouping API: EmbeddedCache.getGroup(G). This
is something useful for us in a broader context, as it is the step needed to be able to
deprecated AtomicMaps and get suggest them being replaced with Grouping.
>
> This approach still has some limitations compared to the current embedded
integration:
> - performance caused by the lack of transactions: this means increased TCP chattiness
between the Hot Rod client and the server.
> - you'd have to handle atomicity, potentially by retrying an operation
>
> What do you think?
>
>
> On Dec 3, 2013, at 3:10 AM, Mircea Markus <mmarkus(a)redhat.com> wrote:
>
>>
>> On Nov 19, 2013, at 10:22 AM, Emmanuel Bernard <emmanuel(a)hibernate.org>
wrote:
>>
>>> It's an interesting approach that would work fine-ish for entities
>>> assuming the Hot Rod client is multi threaded and assuming the client
>>> uses Future to parallelize the calls.
>>
>> The Java Hotrod client is both multithreaded and exposes an Async API.
>>
>>>
>>> But it won't work for associations as we have them designed today.
>>> Each association - or more precisely the query results to go from an
>>> entity A1 to the list of entities B associated to it - is represented by
>>> an AtomicMap.
>>> Each entry in this map does correspond to an entry in the association.
>>>
>>> While we can "guess" the column names and build from the metadata
the
>>> list of composed keys for entities, we cannot do the same for
>>> associations as the key is literally the (composite) id of the
>>> association and we cannot guess that most of the time (we can in very
>>> pathological cases).
>>> We could imagine that we list the association row keys in a special
>>> entry to work around that but this approach is just as problematic and
>>> is conceptually the same.
>>> The only solution would be to lock the whole association for each
>>> operation and I guess impose some versioning / optimistic lock.
>>>
>>> That is not a pattern that scales sufficiently from my experience.
>>
>> I think so too :-)
>>
>>> That's the problem with interconnected data :)
>>>
>>> Emmanuel
>>>
>>> On Mon 2013-11-18 23:05, Mircea Markus wrote:
>>>> Neither the grouping API nor the AtomicMap work over hotrod.
>>>> Between the grouping API and AtomicMap, I think the one that would make
more sense migrating is the grouping API.
>>>> One way or the other, I think the hotrod protocol would require an
enhancement - mind raising a JIRA for that?
>>>> For now I guess you can sacrifice performance and always sending the
entire object across on every update instead of only the deltas?
>>>>
>>>> On Nov 18, 2013, at 9:56 AM, Emmanuel Bernard
<emmanuel(a)hibernate.org> wrote:
>>>>
>>>>> Someone mentioned the grouping API as some sort of alternative to
>>>>> AtomicMap. Maybe we should use that?
>>>>> Note that if we don't have a fine-grained approach we will need
to
>>>>> make sure we *copy* the complex data structure upon reads to mimic
>>>>> proper transaction isolation.
>>>>>
>>>>> On Tue 2013-11-12 15:14, Sanne Grinovero wrote:
>>>>>> On 12 November 2013 14:54, Emmanuel Bernard
<emmanuel(a)hibernate.org> wrote:
>>>>>>> On the transaction side, we can start without them.
>>>>>>
>>>>>> +1 on omitting transactions for now.
>>>>>>
>>>>>> And on the missing AtomicMaps, I hope the Infinispan will want to
implement it?
>>>>>> Would be good to eventually converge on similar featuresets on
remote
>>>>>> vs embedded APIs.
>>>>>>
>>>>>> I know the embedded version relies on batching/transactions, but
I
>>>>>> guess we could obtain a similar effect with some ad-hoc commands
in
>>>>>> Hot Rod?
>>>>>>
>>>>>> Sanne
>>>>>>
>>>>>>>
>>>>>>> On Tue 2013-11-12 14:34, Davide D'Alto wrote:
>>>>>>>> Hi,
>>>>>>>> I'm working on the integration between HotRod and
OGM.
>>>>>>>>
>>>>>>>> We already have a dialect for Inifinispan and I'm
trying to follow the same
>>>>>>>> logic.
>>>>>>>> At the moment I'm having two problems:
>>>>>>>>
>>>>>>>> 1) In the Infinispan dialect we are using the AtomicMap
and the
>>>>>>>> AtomicMapLookup but this classes don't work with the
RemoteCache. Is there
>>>>>>>> an equivalent for HotRod?
>>>>>>>>
>>>>>>>> 2) As far as I know HotRod does not support transactions.
I've found a link
>>>>>>>> to a branch on Mircea repository:
>>>>>>>>
https://github.com/mmarkus/ops_over_hotrod/wiki/Usage-guide
>>>>>>>> Is this something I could/should use?
>>>>>>>>
>>>>>>>> Any help is appreciated.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Davide
>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> infinispan-dev mailing list
>>>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>>>>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> infinispan-dev mailing list
>>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>>>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>> _______________________________________________
>>>>>> infinispan-dev mailing list
>>>>>> infinispan-dev(a)lists.jboss.org
>>>>>>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>> _______________________________________________
>>>>> infinispan-dev mailing list
>>>>> infinispan-dev(a)lists.jboss.org
>>>>>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>
>>>> Cheers,
>>>> --
>>>> Mircea Markus
>>>> Infinispan lead (
www.infinispan.org)
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> infinispan-dev mailing list
>>>> infinispan-dev(a)lists.jboss.org
>>>>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>> _______________________________________________
>>> infinispan-dev mailing list
>>> infinispan-dev(a)lists.jboss.org
>>>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>> Cheers,
>> --
>> Mircea Markus
>> Infinispan lead (
www.infinispan.org)
>>
>>
>>
>>
>
> Cheers,
> --
> Mircea Markus
> Infinispan lead (
www.infinispan.org)
>
>
>
>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
Cheers,
--
Mircea Markus
Infinispan lead (
www.infinispan.org)