Installation and Hardware Setup

There is very simple Clusterpoint Server software installation procedure in Clusterpoint architecture: the complete database platform software, same binary file for for any types of our licenses on a particular operating system (Linux, FreeBSD or MacOS; please contact us for Windows on support@clusterpoint.com), can be downloaded from the Internet from www.clusterpoint.com Download section.  

Clusterpoint DBMS is installed as the same copy of software identically on all cluster hardware servers (cluster nodes).

Clusterpoint Server was designed to operate as transparent cluster software, and each hardware computer dedicated to serve Clusterpoint DBMS database platform, can be installed in the same way across all cluster nodes.  Adding a new node requires just to install another copy of the same software on new hardware.

Typically Clusterpoint software is set up on a dedicated hardware consisting from multiple hardware server nodes as illustrated in the sample network schema below. 

You can use any commodity hardware servers to run Clusterpoint software. 

There is no special requirements against hardware, but large RAM and disk storage space.  Clusterpoint Server software is engineered to utilize all memory capacity which you can provide.  You can also use any commodity networking gear, a dedicated or a shared switch for interconnecting all Clusterpoint Server software installed hardware computers.  In most simple setup like illustrated below, each server is connected to the switch bu a single network link. Clusterpoint Servers communicate among themselves over TCP/IP and UDP networking protocols.  Although you can use public IP addresses to configure your hardware servers, for security reasons it is recommended to use internal or so called NAT addresses for operating Clusterpoint Server database platform cluster software and hardware infrastructure.  It will most likely operate just at the data storage tier, as a database storage for your application and web servers, and public access for administration then is not necessary, you can only access this cluster from your internal network.  Any 3rd party applications can still be allowed to access Clusterpoint Server functionality from Internet through Clusterpoint API, which is based on web http/https protocols and simple XML messaging, where you can securely configure your corporate firewall to enable such access only to your trusted partners or customers.  

Normally all other functional services of Clusterpoint databases will be accessible to end-users only through your application and web services. 

Obtaining Clusterpoint DBMS License

For the first time Clusterpoint DBMS software installation, right after its installation package download, you can freely use a full-fledged Clusterpoint Server software under Clusterpoint Evaluation License for 30-days .  

After that test-driving period, you should make your license registration on our Web site by filling in our simple Registration Form, and obtain your free (for example, Clusterpoint Free Community License) or commercial license file with our custom issued unique license number for your organization or your project, which we send you through your registered email.  This file you should copy into the same directory where Clusterpoint DBMS software was installed, following installation instructions.  The license file will serve as legal proof that you are using properly licensed Clusterpoint DBMS software product.  You can then see the license details in our web-based Clusterpoint Manager application launched from that server (any cluster node can be used). After you launched Clusterpoint Manager, you have a default password protected root administration access and can start configuring and managing all other cluster nodes from easy to use web-based user interface of Clusterpoint Manager.

All servers running Clusterpoint software automatically become "aware" of each other in the corporate network, and automatically show up in our web-based management application Clusterpoint Manager, enabling centralized management environment for all your databases.  Data storage and retrieval, search and indexing tasks are also distributed automatically among cluster nodes by Clusterpoint Server software, according to specific Clustering configuration, which again can be set up and custom configured, using Clusterpoint Manager. 

Clusterpoint Manager Application

You can securely manage all Clusterpoint databases within your organization, using Clusterpoint Manager, an application for centralized management, administration and configuration of all Clusterpoint database infrastructure software system across your entire enterprise.  It has an easy to use web user interface, and you can manage all cluster nodes, Clusterpoint server software, database storages, users and their access rights within cluster of server which you installed with Clusterpoint software.  Clusterpoint Manager is a web-server module, which installs on any cluster node, but can be used from only a single node to administer the whole cluster (there is no single point failure, nor master management node in Clusterpoint architecture).  The management system architecture is shown below.

Centralized administration of all Clusterpoint databases and servers through a Web browser-based interface allows database administrators (DBAs) to configure and manage productively distributed cluster databases and mirrored databases, or a single database on any single cluster node, setting up and configuring any cluster database configurations for as many database storages as necessary, and can be handled by hardware. 

Tho most powerful feature of Clusterpoint Manager is a single-click configuration changes for a Storage Policy or Document Policy, small XML configuration files customizing your database indexing and search preferences.  Those files are located under each named storage directory, and need to be copied into multiple copies for all cluster nodes, handling 1/Nth part of the same name cluster storage. 

 With Clusterpoint Manager you can enjoy a single configuration update operation for cluster databases implemented as a one click update options, which copies any changed configuration files across all cluster nodes for that particular clustered storage. 

There is no need to configure and manage each cluster node separately per each hardware server in the Clusterpoint architecture.

 You can also start, stop all instances of database servers for cluster databases with a cluster wide management commands.  This is really convenient and time-saving facility.

Clusterpoint Manager functionality mirrors many Clusterpoint database concept and design principles, where possible simplifying repetitive, time-consuming database administration tasks.  Please see also section Clusterpoint Technology Overview.

For more about Clusterpoint Manager features please see section Technology / Features. 

Simple File System Organization

Every named Clusterpoint database storage resides under the same name file system folder.  A term 'storage' is used in the same context as 'database' in Clusterpoint architecture, there are no differences.  A storage is just a collection of all your XML documents, which you as our customer want to store into a single searchable database under the unique name.  Storage name is all that is necessary to know to access Clusterpoint database infrastructure functionality.

Depending on cluster configuration set up in Clusterpoint Manager, Clusterpoint Server software treats all same name databases on each cluster node either as a multiple mirrored database copies, or as a striped (distributed) cluster databases with their parts distributed across all cluster nodes.  All the data for the same name storages will be always under the same identically named directories per each cluster node, no matter what cluster configuration is configured: for database striping or clustering (please see also section Clustering).   

Each named storage is completely autonomous directory containing the following files: all data files (storing copies of all original customer XML documents as Clusterpoint Server document Repository), all index files (storing Clusterpoint Index), all configuration files (Document Policy, Storage Policy and others), all transaction log files and all error log files for a particular storage.  You can copy, swap, backup and restore those storages as a simple whole directory copy or move commands, easily delete in mass any unnecessary storages across the whole cluster, operatively replace malfunctioning hardware by temporarily copying all storages on a neighbor cluster node or spare disk, and copy storage data back when hardware was replaced, to quickly restore previous quality of database services etc.  For example, sysdmins can simply copy same name Clusterpoint storage directories from one server to another for any routine equipment upgrade or maintenance task, using just very basic levels of file management skills.  There is no need to learn database structure either. 

This simplicity of Clusterpoint database storage organization on a file system level brings multiple benefits, making system administration and servicing understandable and enabling fast problem fixing with basic system administration operations even on a file system level.  We decided that ease of quick problem fixing is much more important management feature in our massively clustered database architecture, than spending time to develop custom sophisticated database file location strategies and then trying to remember all that complexity and manage it.  Cluster nodes in most very large database use cases will be storing just a small part of the whole database, and organizing all database files per same name directory makes better chances to quickly fix any unforeseen problems in cluster systems.  Most likely problems in massively clustered environment will be cased by hardware failures and the quick replacement of a cluster node hardware should be a priority.  This our very simple and uniform principle of location of database files per each cluster node also is making job duties more convenient and less error-prone for both DBAs and sysadmins. 

Clusterpoint Manager Functionality

Please see section Features / Management features. 

Clustering Controlled at DBMS Level

From the application developer point of view there is no need to learn about specific hardware configuration of cluster nodes in Clusterpoint architecture, no need to know number of hardware servers in a cluster, no need to know how many distributed parts of a database exists per single named cluster storage, no need to know their exact location within a cluster per server etc.  The application developers even do not need to address specific custom databases indexing and search logic into their application software.

Customer applications can simply address any Clusterpoint database as a single logical named database entity, using our Clusterpoint API. 

All underlying complexity of cluster is being handled by Clusterpoint Manager:  hardware configuration, database clustering methods, configuration of a particular database indexing and search policy, operation and control of cluster databases in Clusterpoint Architecture.   It is a centralized management system, which only DBAs or data center system administrators can access and use.  

This effectively also partitions database administration tasks separated from application development and database design tasks, when databases are run into production environment. Application software developers normally can not access a production database due to security reasons.  Yet, our multiple administration account management system allows application developers to freely access and administer development databases, while performing software testing and the design of the corporate database search ranking algorithms.  Afterwards, only resulting Document Policy and Storage Configuration files are being transferred through simple copy-paste operations (remember, they are are just small XML configuration files) by DBAs to production databases, to make required changes into Indexing and Search algorithms, for example, to provide some new application functionality for the same customer XML database, using the latest software versions delivered from application developers for a production database.  Please also read about Clusterpoint Indexing system, Clusterpoint Search and Clusterpoint Information Ranking - those key features were designed to eliminate the need to address indexing and search logic into application software in Clusterpoint architecture.

Clusterpoint Management system was designed to support this our approach to make application software development as much as possible database structure agnostic, make it more simple to develop and understand, and radically speed software development time up.  With Clusterpoint there is much less, if any, database partitioning logic at application software level needed.  Major efforts are still required to develop user interface data entry and reporting forms from software developers, as those parts will always be application database structure specific, but when done, all the rest of software development (indexing logic, search logic, database storage partitioning, file location specific partitioning in software etc.) become largely irrelevant from application developers point of view and does not require to address in application software or require only minimum efforts.  These tasks can be done by easily configurable and customizable Information Ranking options, applicable to customer databases directly from Clusterpoint Manager or through a command line text editor.

Clusterpoint Manager also allows to ad new cluster nodes, to scale application software performance and database storage capacity 'behind the scenes'.  Again, this can be done on-the-fly for any Clusterpoint database, using Clusterpoint Manager utility, by DBAs.  Application software developers should not worry about capacity problems, nor performance problems, if expendable spare hardware can be re-configured to increase total capacity of cluster infrastructure.  For example, efficient database server work for double number of simultaneous search users can be easily restored, by doubling number of servers, and mirroring the existing cluster database to newly added servers. Not a single line of application software need to be changed.  For DBAs its also a quick job: install Clusterpoint software on extra hardware and use Clusterpoint Manager to expand and configure existing cluster databases to new capacity, then configure existing web server or an application server load balancing system to pool mirrored servers.  Same easiness is for expanding cluster database storage capacity.  Any cluster database updates will be automatically re-balanced by Clusterpoint Server,  just right after cluster configuration changes, when the new cluster node will be configured to expand existing storage from N nodes to N+1 node through Clusterpoint Manager.  Please see more about Clustering. 

Cluster-wide User Management

There are the following basic user management features built-into in Clusterpoint Server database software:

• password protected user accounts: commonly other API services, as Clusterpoint platform is typically used as data storage engine
• groups of users: to group access to specific storages and their access rights per organization or task
• API level command-specific access rights: up to very specific API commands
• multiple administrator accounts: user accounts with their own rights to see, manage, change or read storages
• audit logs: for all Clusterpoint Manager activities
• directory services: optional plug-in to connect to LDAP or ADS for user authentication
• 'root' administrator account: the first default user account, used to set up and manage all others

Users per cluster storages are set up and managed centrally, with a one click changes to access rights propagated to all cluster nodes handling the same name storage parts.  Use Clusterpoint Manager to provide this functionality.

You can run any Clusterpoint database with different user accounts, access rights and on different operating system platforms, for example, under a virtual machine or a virtual appliance software environment, and still manage all users enterprise-wide, using Clusterpoint Manager. 

Support for SNMP Tools

We support standard enterprise system management and monitoring tools, using SNMP protocol and standards, which are supported by all major enterprise software systems.  

As Clusterpoint Server hardware cluster nodes are normally installed on dedicated servers (like an appliances), we provide means to monitor them remotely also from 3rd party systems.

Any common enterprise system management tool such as HP OpenView (www.hp.com) or open source Zabix (www.zabix.com), through our plug-in SNMP module, can view and monitor all Clusterpoint Servers for their system operational parameters to detect or notice problems in timely manner.

We provide the following information through our SNMP module:

• appliance status (a hardware server making a cluster node)
• free disk space
• free memory
• CPU workload
• number of storages
• documents per each storage
• other parameters what our API command 'status' returns

Customers can monitor well-being of Clusterpoint database infrastructure servers in their own preferred management environment, and in case of problems log in to Clusterpoint Manager for more specific problem resolution, transaction tracking or specific database management activity.