Rating processes in Singl.eView in Detail:
Basically there are five billing components that execute the rating functions in Singl.eView mentioned as below:
Event Rating Broker (BKR) process
Event Normalisation (ENM) process
Event Rating (ERT) process
Event Rating Output (ERO) process
TRE Rating Server (trerate)
The combination of the BKR, ENM, ERT, and ERO processes is called as Rating engine. The trerate server assists in real-time rating by providing EPM functions for the calculation of real-time rating charges and account balances.
Event Rating Broker (BKR)
The BKR process manages the rating engine by controlling the ENM,
ERT, and ERO child processes, and creates in shared memory the interprocess
communication queues, service cache, customer cache, account
cache, and rating subtotal cache. Additionally, an event cache is created for each BKR process.
Depending upon configuration, the BKR process creates upon startup the
inter-process communication mechanisms for the ENM, ERT, and ERO
processes. The BKR process then starts all ENM, ERT, and ERO
processes that are configured in that way.
All ENM, ERT and ERO processes associated with a given BKR must
run on the same node or Convergent Billing instance as the BKR.
Inter-process mechanisms created by the BKR are used to transmit:
· Normalized events from ENM processes to ERT processes for rating.
· Normalised events, erred events, charges, and details of normalized event files to an ERO process for output. Multiple BKR processes can exist for each Convergent Billing instance.
Event Normalisation (ENM)
The ENM process is responsible for converting incoming events (which
may include usage records or events that represent either recurring time
periods or changes in product sets) into a standard, or normalised, format.
Depending on the ENM configuration, it either waits for a command to
start normalising events from a binary file on disk containing events of a
specified type, or listens on a socket or a named pipe for one or more realtime
rating events from the TRE Rating Server. Regardless of the input
method, the source is represented by a normalised event file and is a
collection of raw events.
The ENM uses both a DIL specification and decoding expressions to
parse, clean, validate, and build normalised events
Raw events are parsed and cleaned into fields based on the DIL (Data
Interface Language) specification. Parsed field values are
assigned to event direct variables.
A sequence of decoding expressions, written in the EPM language, allows
additional validation to be performed on event direct variables, and
allowing new information to be derived from the information in the raw
event and assigned to event direct variables.
For example, a table lookup can be used to map a Cell ID from the raw event into a Region Code, and store the Region Code in an event direct variable.
The normalised event is built by ordering the event direct variables
according to the record layout. The ENM allocates each new event a
unique ID that spans reprocessing.
Event Rating (ERT)
The ERT process is responsible for applying tariffs to the normalised
events to generate charges. The ERT uses data stored in the service cache
to determine all services associated with each normalised event.
The service cache data also includes products associated with each
service. Using the associated products, the ERT process determines the
associated tariffs and applies prioritization rules to arrive at a final set of
product tariffs.
Each tariff, if eligible, is evaluated and can:
Generate a number of charges.
Update the unbilled account balance of an account.
Update ratings subtotals and clear reservations.
The ERT process writes its updates to the account and rating subtotal
caches and writes the generated charges to the event cache. Updates are
committed to the database when the ENM process is synchronized with
the ERO process.
Charges generated by the ERT process are transferred to an ERO process
using inter-process communication queues.
Event Rating Output (ERO)
The ERO process takes normalised events from the ENM and charges
from the ERT, and sends them to the database. Database updating can be
done directly.
During rating, each ENM establishes a link with one ERO (if there are
multiple, running ENM or ERO processes), and maintains the link only
for the current event file. The ENM selects an ERO that is not currently
processing records, based on the ERO’s priority (set up in the process
configuration).
The ENM-ERO pair communicates with each other (using the interprocess
communication shared memory segments) to ensure that events
pass successfully through the ENM and one or more ERTs, and that the
ERO remains synchronised with the ENM.
Additionally, the ERO is responsible for ensuring that the account and
rating subtotal caches are kept in synchronisation with the database. The
ERO commits account and rating subtotal cache updates to the database
when synchronising with an ENM process.
Events that are successfully normalised and rated are passed to the
database normalised event and charge tables, either immediately or by
loader file. Normalisation or rating error events (collectively referred to
as error events) are stored in the database.
Rating Server (trerate)
The TRE rating server (trerate) provides a single bi-directional
interface between a remote host and the rating engine.
The rating server provides an interface to callback functions that can be
used to query charge calculations and manage customer accounts
(biEventRateAndReturn). The reservation interface and external
access to subtotals, unbilled amounts, and all read-only information are
provided by trerate-registered functions.
trerate also handles asynchronous functions executed for the
replication of cache changes (in a high availability configuration) and
provides functions that send completed raw events to the rating process
(biEventRateAndStore).
Conclusion
Event rating is the process in which input events are converted into rated
events, using tariffs that have been defined for each product. Events can
be rated either in batches or in real time. This document helps the reader to understand the flow of the process involved in the rating. Please note that the Real-time rating with authorisations is made possible by the Commerce Engine using the TRE Rating Server (trerate), thus giving excellent billing solutions to the Telecom companies using Singl.eView.
Friday, January 8, 2010
Singl.eView Rating processes in detail
Subscribe to:
Post Comments (Atom)
3 comments on "Singl.eView Rating processes in detail"
Hey
Thanks for sharing this! SV is overwhelming if you don't know what happens where! Keep sharing!
Thanks for sharing such valuable information on Single View Application.
Thank you, this is very helpful information
Post a Comment