MarcL: Data vaulting and document management are critical functionality in PLM, can you describe how the managed data is stored in a distributed environment with heterogeneous file types, and explain the scenarios for more complex distributed meta-data servers in large scale global environments?
Peter Schroer:
Aras Innovator enables 3 Vault Server configurations out-of-the-box
- single server with a single master vault
- distributed vault servers… each File is located in a single vault within a network of vault servers
- replicated vault servers, each File is stored in one or more Vault Servers simultaneously throughout the network
In all 3 configurations, the physical arrangement of the vault servers is invisible to the end-users.
Regardless of where the File is physically stored, the entire configuration of vault servers appears as a single logical vault to the end users. Users search for and view documents without considering which vault server is streaming the files.
The administrator controls allocation of Files to Vaults, and has simple tools for specifying which Files should be replicated to which destinations, when (immediate or periodically based on configurable rules), under which circumstances (execution of business logic, lifecycle transitions, etc).
There are also options for Cloud replication through Aras partner Ilesfay.
Distribution meta-data for failure / redundancy is normally implemented using standard Microsoft tools for clustering of servers, in which case the replicated meta-data store is always a 100% copy.
If the reason for distribution of meta-data is performance or access to data over a sub-optimal WAN connection, then the Aras recommendation is to install separate instances of Aras Innovator, and use Federation techniques to distribute data between the systems.
If a 100% copy is desired then a master-slave arrangement using standard Microsoft SQL Server replication is recommended.
If there will be editing in all instances of a distributed PLM system or only sub-sets of the data catalog are desired in one location, then a publish mechanism is recommended. The core Aras Innovator framework supports migration of data via XML export / import, and the Admin must develop the rules for which product information is extracted and at what frequency and how any conflicts that might occur will be resolved.
It’s also important to understand that Aras Innovator can Vault any file-based data regardless of format, as well as, being able to store any data representation in the database as BLOBs (Binary Large OBjects) if required. Both Vaulted and BLOB data can be stored in the repository and used in PLM processes, configurations, versioned, access controlled, etc.
Image from Razorleaf blog
Slides: ACE 2010 Workshop – Aras Vault Replication
Video: Aras Partner Ilesfay Point-to-Cloud Replication Option