|
Editors Top Solution
|
| Title
- HP ITSM |
| Author
- Hewlett-Packard |
|
Short Description
|
| HP IT
Service Management ITSM combines powerful HP OpenView software
with years of experience to transform your IT into a real business
and competitive differentiator. See how HP IT Service Management
solutions bring people, processes, and technology together to
capitalize on change. |
| HP ITSM
Download |
|
| - Emic acluster |
| Emic acluster
Description |
Emic a/cluster provides high availability, scalability and manageability services for Apache. a/cluster keeps your website up and running despite any outages of individual web servers.
The architecture can be divided in two major parts: IP-layer clustering that enables the cluster to appear as a single server, and the application data synchronization logic that provides the synchronous data replication between cluster nodes.
All cluster nodes share the same IP address, which makes the whole cluster appear as a single IP network element. The cluster dynamically selects one of the cluster nodes as an active dispatcher node. This node receives all incoming queries that are sent to the cluster, and then, routes them to other cluster nodes based on the load balancing rules. With this shared IP clustering technique, the whole cluster appears as a one logical server to the application.
The application data synchronization layer handles the actual database logic. It is a shared-nothing architecture.
This means that all cluster nodes have a full copy of the entire database, which is kept in sync between all nodes. The synchronization is based on reliable group communication mechanism that guarantees that messages are either delivered in the same sequence to all nodes in the cluster or not at all. This is called multi-master database replication.
Once a client connects and executes a query, the application data synchronization logic intercepts the query, and based on the nature of it does one of the following. If the query is a read, it is executed against the local database and the result is reported back to the client. If the query is a write or is otherwise modifying data, the query processed on the cluster node, but will wait for the data synchronization process.
After this process is complete the write transaction will have been replicated to all cluster nodes and is applied to the database on all nodes, and the result is returned to the client. All queries are ordered to guarantee that once a data-altering query is finished, the result is visible in all cluster nodes at the same time. The application will always see the correct data regardless of which node has processed the query. |
| - Emic acluster
Download |
|
