On Fri, Mar 20, 2009 at 7:10 PM, Orlando Andico <[email protected]> wrote: > > 1) has anyone really looked into MySQL Cluster? > > I have. That shared-nothing implementation is.. really hard to use in > practice. You are limited to available memory (!) because NDB Cluster > is an in-memory implementation. And look at their references.. who has > heard of Zillow.com?
if you said about rac that people dont know how to tune or configure rac.. they same true also to people who dont know how to configure mysql cluster... alcatel uses mysql cluster for their home location register (HLR).. HLR is a central database that contains details of each mobile phone subscriber that is authorized to use the gsm network.. vodafone in UK uses mysql cluster too.. > I looked into MySQL Cluster before at a prior role, and I concluded > that for what we wanted, MySQL Cluster was more trouble than it was > worth. My understanding is that due to the replication topology, your > writes will not scale well as the number of nodes increases as you > have to propagate changes OVER THE NETWORK to the other nodes. > > So much for linear scalability... the mysql cluster replication provides highly availability in case one node fails the other node will take care of that data... usually replication took place when an update on the master server... it propagates to a slave server and not to the rest of the nodes... > MySQL Cluster's overview claims "few milliseconds" query latency and > "tens of thousands" of transactions per second. I'd bet those are > mostly reads. > > Oracle In-Memory Database can do SIX MICROSECOND selects and FIFTEEN > MICROSECOND updates, and up to 750K selects/second (write performance > is considerably less, around 200K/second). Granted this is not Oracle > DB, but it's the more logical comparison to MySQL Cluster. > > > > > 2) Amazon has the world's largest data warehouse. And it's not running > a shared-nothing MySQL architecture. i didnt mention amazon uses mysql ... what i mentioned was that amazon used a shared-nothing architecture also just like google for scalability reason as oracle rac was well known of slow i/o performance (eg. weaknesses in parallel query planning.. inefficient movement of data between nodes for execution of complex joins and aggregations) thus exadata storage exists... fooler. _________________________________________________ Philippine Linux Users' Group (PLUG) Mailing List http://lists.linux.org.ph/mailman/listinfo/plug Searchable Archives: http://archives.free.net.ph
