Cassandra is not being used to generate the Twitter identifiers. Twitter, like most places using Cassandra, has more than one database system in production.
UUIDs are not at risk of conflicts with billions of rows. b On Thu, Mar 25, 2010 at 5:57 AM, Jaepil Jeong <zgdr...@gmail.com> wrote: > Hi there, > Thanks to all for reply. But I still have a question: > 1. When I using Twtiter via Tweetie which is iPhone application, I can see > the unique ID for each of users in their personal profile page. It seems > like incremental number. As far as I know, Twitter using Cassandra for its > back-end. My question is how do they do that? > 2. I know UUID is the best solution to generate unique ID. However, I worry > about value conflicts with billions of columns. Is there anyone who has > experience about this? > Thanks, > > > On Thu, Mar 25, 2010 at 10:37 AM, Muhammed Nasrullah <nasrul...@gmail.com> > wrote: >> >> You replace it with an UUID. In a true scalable distributed system, you >> should not have an auto_increment. If you are writing to 10 >> nodes simultaneously, it becomes near impossible to keep a single >> incrementing value being used by the entire system without causing a lot of >> write contention. >> This is how you generate a Time-UUID in >> Java: http://wiki.apache.org/cassandra/FAQ#working_with_timeuuid_in_java >> >> On Wed, Mar 24, 2010 at 9:57 PM, Jaepil Jeong <zgdr...@gmail.com> wrote: >>> >>> Hi there, >>> >>> I just started research about Cassandra to replace MySQL, and I have a >>> question: How can I replace the "auto increament" attribute in MySQL >>> with Cassandra? If I can't, how can I generate an ID which is globally >>> unique for each of columns? >>> >>> Thanks, >>> >>> Sent from my iPhone >> > > > > -- > Jaepil Jeong > Software Architect > > twitter: http://twitter.com/JaepilJeong > msn: zgdr...@hotmail.com > > Everything arises, everything falls away. > >
