- Based on SMP technology combined with a communication network connecting the SMP systems to
form a MPP system
- All processors in a same SMP node are connected by a virtual BYNET.
- The Teradata Database doesn’t rely on the hardware platform for parallelism, scalability,
reliability, or availability.
- A single processor node is connected to one or more disk arrays with the following installed on the
node:
o Database software
o Client interface software
o Operating system
o Multiple processors with shared-memory
o Failsafe power provisions
o Hardware inter-processor network to link nodes on an MPP system.
o BYNET provides a node-to-node data transfer bandwidth of 120 MB/sec and can linearly scale, supporting up to 512 nodes on a single system.
Parallel Database Extensions (PDE) software provides the ability to:
• Execute vprocs
• Run the Teradata RDBMS in a parallel environment
You can start, reset, and stop the PDE on Windows systems using the Teradata MultiTool utility and on UNIX MP-RAS systems using the xctl utility.
- There are two types of vprocs:
§ PE (Parse Engine)
§ AMP (Access Module Processor)
- The PE performs session control and dispatching tasks as well as parsing functions
- The Parsing Engine (PE) is the vproc that communicates with the client system on one side and with the AMPs (via the BYNET) on the other side.
- Each PE executes the database software that manages sessions, decomposes SQL statements into steps, possibly parallel, and returns the answer rows to the requesting client.
- The PE software consists of the following elements:
o Parser
o Optimizer
o Generator
o Dispatcher
o Session Control
- Heart of the Teradata RDBMS.
- Retrieve and update data on the virtual disks (vdisks)
- A single system can support a maximum of 16,384 vprocs.
- The maximum number of vprocs per node can be as high as 128.
- AMPs control the management of the Teradata RDBMS and the disk subsystem, with each AMP
being assigned to a virtual disk (vdisk).
- Each AMP provides the following set of functions:
o Database management
§ Locking
§ Joining
§ Sorting
§ Aggregation
§ Output data conversion
§ Disk space management
§ Accounting
§ Journaling
o File-system and disk management
§ Each AMP manages a portion of the physical disk space.
§ Each AMP stores its portion of each database table within that disk space.
Both hardware and software provide fault tolerance, some of which is mandatory and some of which is optional.
o Vproc migration
o Fallback tables
o AMP clusters
o Journaling
o Archive/Recovery
o Table Rebuild Utility
o Multiple BYNETs
o RAID disk units
o Multiple channel and LAN connections
o Isolation from client hardware defects
o Battery backup
o Redundant power supplies and fans
o Hot swap capability for node components
o Cliques
This comment has been removed by the author.
ReplyDelete