Those Are Not Transactions (Cassandra 2.0)

#1

Post by djbrowning » Fri, Sep 06, 2013, 3:51 PM UTC

Those Are Not Transactions (Cassandra 2.0)
blog.foundationdb.com

Re: Those Are Not Transactions (Cassandra 2.0)

#2

Post by voidmain » Fri, Sep 06, 2013, 4:08 PM UTC

I'm the author of the post, if anyone has any questions (or corrections :-)

Re: Those Are Not Transactions (Cassandra 2.0)

#3

Post by misframer » Fri, Sep 06, 2013, 11:10 PM UTC

voidmain wrote: post #2

I'm the author of the post, if anyone has any questions (or corrections :-)

)

:)

Re: Those Are Not Transactions (Cassandra 2.0)

#4

Post by xaprb » Sat, Sep 07, 2013, 12:53 AM UTC

voidmain wrote: post #2

I'm the author of the post, if anyone has any questions (or corrections :-)

I agree -- I dislike when important technical concepts are redefined for marketing purposes. It's a bit like lightweight marital fidelity, where your spouse is faithful to you in specific circumstances.

Re: Those Are Not Transactions (Cassandra 2.0)

#5

Post by thesnider » Sat, Sep 07, 2013, 4:08 AM UTC

voidmain wrote: post #2

I'm the author of the post, if anyone has any questions (or corrections :-)

While your analysis of Cassandra's Paxos implementation may be correct (I haven't checked), I've never heard of anyone else (especially not Google) using a model where there is only one Paxos server per datacenter -- usually there are many more.

Re: Those Are Not Transactions (Cassandra 2.0)

#6

Post by penguindev » Tue, Sep 10, 2013, 8:20 PM UTC

Yes, 'Heavyweight CAS' would be more appropriate.