Re: ZAB kick Paxos butt?
Yeah, actually I have no doubts about Paxos protocol itself but rather the state machine implementation part (as described in Paxos made simple,section 3) where there could be multiple Paxos instances. shouldn't the Paxos instance execution be serialized in order to make the state machine abstraction useful/friendly for the real world use? if one paxo instance fails application will be notified so that corresponding actions could be taken(retry,rollback,notify client...etc), instead of blindly continuing and getting unpredictable results later on. Actually in the Google Chubby case, the database changelog is being streamed into the Paxos cluster,how can they afford to lose some of the logs without breaking the database integrity? Did I miss something? On the other hand, I think adopting the FIFO based protocol is a very smart engineering decision. It makes the whole thing less complicated and is also more powerful. E.g. it saves you guys the efforts to invent another language/compiler(like the Google ppl does). Just curious, giving how persuasive the TCP stack is deployed today, why the research community still need to stick to the asynchronous system assumption? Just because TCP sounds uncool than asynchronous system on paper? hehe.. Cheers Qing >
Re: ZAB kick Paxos butt?
Qing -
Also, as you pointed out, ZAB requires this FIFO property of the
point-to-point links. Paxos copes with more adversarial networks which allow
reordering and missed messages. It's easy to alter Paxos so as not to
'publish' the results of consensus rounds where there are gaps in the
previous commit history. (You may be interested in the 'Fast Paxos' paper by
Lamport which talks about making the protocol 2-message optimal in all cases
when order is not important, i.e. the messages commute). You can express the
ordering dependency between messages by supplying a proposal number with
each that is monotonically increasing in causal order.
ZAB takes care of all of this for you by using TCP sequence numbers and
getting the deep pipelining available by knowing that there are no updates
being voted on depend on updates that have not yet arrived, the 'cost' is
relying on a stronger network model than Paxos presupposes.
Henry
2010/1/20 Benjamin Reed
> hi Qing,
>
> i'm glad you like the page and Zab.
>
> yes, we are very familiar with Paxos. that page is meant to show a weakness
> of Paxos and a design point for Zab. it is not to say Paxos is not useful.
> Paxos is used in the real world in production systems. sometimes there are
> not order dependencies between messages, so Paxos is fine.
>
> in cases where order is important, multiple messages are batched into a
> single operation and only one operation is outstanding at a time. (i believe
> that this is what Chubby does, for example.) this is the solution you allude
> to: wait for 27 to commit before 28 is issued.
>
> for ZooKeeper we do have order dependencies and we wanted to have multiple
> operations in progress at various stages of the pipeline to allow us to
> lower latencies as well as increase our bandwidth utilization, which led us
> to Zab.
>
> ben
>
>
> Qing Yan wrote:
>
>> Hello,
>>Anyone familer with Paxos protocol here?
>>I was doing some comparision of ZAB vs Paxos... first of all, ZAB's
>> FIFO
>> based protocol is really cool!
>>
>> http://wiki.apache.org/hadoop/ZooKeeper/PaxosRun mentioned the
>> inconsistency case for Paxos("the state change B depends upon A, but A was
>> not committed").
>> In the "Paxos made simple" paper, author suggests fill the GAP (lost
>> state
>> machine changes) with "NO OP" opeartion.
>>
>> Now I have some serious doubts how could Paxos be any useful in the real
>> world. yeah you do reach the consesus - albeit the content
>> is inconsistent/corrupted !?
>>
>> E.g. on the wiki page, why the Paxos state machine allow fire off 27,28
>> concurrently where there is actually depedency? Shouldn't you wait
>> instance
>> 27 to be committed before start 28?
>> Did I miss something?
>>
>> Thanks for the enlight!
>>
>> Cheers
>>
>>Qing
>>
>>
>
>
Re: ZAB kick Paxos butt?
hi Qing,
i'm glad you like the page and Zab.
yes, we are very familiar with Paxos. that page is meant to show a
weakness of Paxos and a design point for Zab. it is not to say Paxos is
not useful. Paxos is used in the real world in production systems.
sometimes there are not order dependencies between messages, so Paxos is
fine.
in cases where order is important, multiple messages are batched into a
single operation and only one operation is outstanding at a time. (i
believe that this is what Chubby does, for example.) this is the
solution you allude to: wait for 27 to commit before 28 is issued.
for ZooKeeper we do have order dependencies and we wanted to have
multiple operations in progress at various stages of the pipeline to
allow us to lower latencies as well as increase our bandwidth
utilization, which led us to Zab.
ben
Qing Yan wrote:
Hello,
Anyone familer with Paxos protocol here?
I was doing some comparision of ZAB vs Paxos... first of all, ZAB's FIFO
based protocol is really cool!
http://wiki.apache.org/hadoop/ZooKeeper/PaxosRun mentioned the
inconsistency case for Paxos("the state change B depends upon A, but A was
not committed").
In the "Paxos made simple" paper, author suggests fill the GAP (lost state
machine changes) with "NO OP" opeartion.
Now I have some serious doubts how could Paxos be any useful in the real
world. yeah you do reach the consesus - albeit the content
is inconsistent/corrupted !?
E.g. on the wiki page, why the Paxos state machine allow fire off 27,28
concurrently where there is actually depedency? Shouldn't you wait instance
27 to be committed before start 28?
Did I miss something?
Thanks for the enlight!
Cheers
Qing
ZAB kick Paxos butt?
Hello,
Anyone familer with Paxos protocol here?
I was doing some comparision of ZAB vs Paxos... first of all, ZAB's FIFO
based protocol is really cool!
http://wiki.apache.org/hadoop/ZooKeeper/PaxosRun mentioned the
inconsistency case for Paxos("the state change B depends upon A, but A was
not committed").
In the "Paxos made simple" paper, author suggests fill the GAP (lost state
machine changes) with "NO OP" opeartion.
Now I have some serious doubts how could Paxos be any useful in the real
world. yeah you do reach the consesus - albeit the content
is inconsistent/corrupted !?
E.g. on the wiki page, why the Paxos state machine allow fire off 27,28
concurrently where there is actually depedency? Shouldn't you wait instance
27 to be committed before start 28?
Did I miss something?
Thanks for the enlight!
Cheers
Qing
