If you want High Availability for free, you will have to look into shared nothing architectures. You have open source solutions such as Sequoia (http://sequoia.continuent.org) that are database neutral that can do the job.
Live migration is doable with shared nothing but it will take longer (basically depends on the size of your VM size).
The post surprisingly does not mention Aster Data Systems which is the datawarehouse behind MySpace. When web sites start to store and analyze every single user click then you quickly get into massive amount of data. It's no surprise that the Petabyte barrier is reached especially with the density of storage increasing at constant cost.
Clustering and High-Availability aspects are not mentioned at all.
MySQL speed will really depend on the database engine you use (MyISAM or InnoDB do not perform the same!). PostgreSQL performance is pretty much consistent across platforms.
Sequoia provides both load balancing and failover. It supports all MySQL table types (and also other databases). One nice feature is transparent failover that completely hides any failure (connection, controller, backend) even in a middle of a transaction.
Note that Sequoia now has a native implementation of the MySQL client library to allow any other API besides JDBC.
For other databases, ODBC support is also available in the Carob project.
I am working for the Sequoia project. I don't think that this post was made by anyone of our group. I don't necessarily share your vision with the advantage of a shared disk architecture over shared nothing.
As a SAN has to be shared between all nodes in your cluster, this already limit the availability and scalability of your database to your SAN capabilities (you are shifting the problem to the disk). With a shared nothing architecture you are also replicating the disks and thus distributing the IO workload as well. Note that when you want to synchronize nodes that are not collocated, then a SAN does not work anymore.
You should read some papers about C-JDBC (the original name of the Sequoia technology) and you will find that the technology can scale quite well. We were able to achieve throughput of a million transaction a minute... that should be quite enough for many businesses.
This looks like a religious war between shared-disk and shared-nothing solutions but a SAN and its admin cost is usually not compatible with someone seeking for an open source solution. If you can afford the SAN, why not just using Oracle RAC?
Slashdot Mirror
If you want High Availability for free, you will have to look into shared nothing architectures. You have open source solutions such as Sequoia (http://sequoia.continuent.org) that are database neutral that can do the job. Live migration is doable with shared nothing but it will take longer (basically depends on the size of your VM size).
The post surprisingly does not mention Aster Data Systems which is the datawarehouse behind MySpace. When web sites start to store and analyze every single user click then you quickly get into massive amount of data. It's no surprise that the Petabyte barrier is reached especially with the density of storage increasing at constant cost.
MySQL speed will really depend on the database engine you use (MyISAM or InnoDB do not perform the same!). PostgreSQL performance is pretty much consistent across platforms.
On the HA side, PostgreSQL has maybe less options: Slony/I (http://gborg.postgresql.org/project/slony1/) for master/slave or Sequoia (http://sequoia.continuent.org/) for multi-master.o n.html) for master/slave, MySQL cluster (http://dev.mysql.com/doc/refman/5.0/en/mysql-clus ter.html) for those who want to switch to a new storage engine (NDB) or Sequoia (URL:http://sequoia.continuent.org/) for multi-master with transparent failover.
MySQL offers MySQL replication (http://dev.mysql.com/doc/refman/5.0/en/replicati
Sequoia provides both load balancing and failover. It supports all MySQL table types (and also other databases). One nice feature is transparent failover that completely hides any failure (connection, controller, backend) even in a middle of a transaction.
Note that Sequoia now has a native implementation of the MySQL client library to allow any other API besides JDBC.
For other databases, ODBC support is also available in the Carob project.
As a SAN has to be shared between all nodes in your cluster, this already limit the availability and scalability of your database to your SAN capabilities (you are shifting the problem to the disk). With a shared nothing architecture you are also replicating the disks and thus distributing the IO workload as well. Note that when you want to synchronize nodes that are not collocated, then a SAN does not work anymore.
You should read some papers about C-JDBC (the original name of the Sequoia technology) and you will find that the technology can scale quite well. We were able to achieve throughput of a million transaction a minute... that should be quite enough for many businesses.
This looks like a religious war between shared-disk and shared-nothing solutions but a SAN and its admin cost is usually not compatible with someone seeking for an open source solution. If you can afford the SAN, why not just using Oracle RAC?