About Broadcast System Information [Under
UMTS]
» RRC -
25.331 «
In last article, we looked at cell search procedure performed during UE power on. Next step would be network registration. NAS/MM entity is responsible for network registration procedures. Note that registration is to be done for both CS and PS domains. Registration can be done separately for CS and PS or can be combined (in which case interface between MSC and SGSN is to be present). CS domain registration is nothing but location update procedure and for PS domain, it is done with GPRS attach procedure. For doing combined CS and PS attach, GPRS attach procedure will be performed - the difference being: "Attach type" in Attach Request would be "Combined GPRS/IMSI attach" rather than "GPRS Attach" or "GPRS attach while IMSI attached". These NAS procedures are similar to what is done for GSM and GPRS (briefed in GSM and GPRS articles).
For NAS procedures, RR connection is needed. RR connection is requested from RRC.
To establish a connection, RRC need to send "RRC Connection Request" over CCCH. During UE power on, transport channel that would be used is RACH. The physical channel would be PRACH. PHY has to perform "random access procedure" at this point in time. Before we look at this procedure, we need to look at RRC states. As overall control of AS is with RRC, RRC states determine channel mapping and various other procedures. To complete our understanding, we will also look at System information. Note that RRC had already used partial system information during cell search for determining PLMN, cell selection etc.
System information comes in System Information Blocks (SIB) (which itself is part of RRC SYSTEM INFORMATION message). SIBs can be segmented or segments can be packed in one message. One SIB cater to particular type of information like SIB type 4 will have Cell identity, Cell selection and re-selection information etc. There are 18 types of SIBs ( 1 till 7 and 11 till 18 - 8/9/10 are not applicable to R5). SIBs are broadcast regularly, but with different repetition rate. System information is actually organised as tree with Master Information Black (MIB) as root. MIB contain references to SIB and (1 or 2) Scheduling blocks (SB). Scheduling blocks contain references to other SIBs as shown below:
Why do we need Scheduling blocks ?
Once UE has camped on a network and read all required system information, it may not be needed to read this information again and again. In fact, for registered UEs, any change in system information is indicated by network in paging message or RRC SYSTEM INFORMATION CHANGE INDICATION message sent on S-CCPCH/FACH channel (depending on RRC state). This way, UE power is saved. In addition, by having tree structure and scheduling blocks for system information, UE need not listen to P-CCPCH, but can (sleep and) listen to (changed or required) SIB at scheduled time. This way, further power saving is achieved.
As MIB is the root, MIB is scheduled at fixed intervals. During cell search, UE must first look for MIB and then proceed accordingly.
In next article we will continue with RRC states.
References: UMTS by Sanchez and Thioune, and WCDMA for UMTS by Holma and Toskala.
| Copyright © Samir Amberkar 2010 | § |
| UE power on - PHY « | UMTS Index | » RRC states |