User Manual

Chapter 3 - Operation and Administration of the Macro BTS NPU Configuration
4Motion 382 System Manual
3.4.12.14 Managing the ASN-GW Keep-Alive Functionality
Once an MS enters the network, its context is stored in ASN entities (BS,
ASN-GW). Dynamically, MS context could be transferred/updated (during HO and
re-authentication) to other entities or duplicated to other entities (separation
between anchor functions such as Authenticator, Data Path and Relay Data
Path).
In certain cases, such as entity reset, other entities are not aware of service
termination of an MS in that entity, and keep maintaining the MS context. This
may result in service failure, excessive consumption of memory resources and
accounting mistakes.
The keep-alive mechanism should be used to clear MS context from all network
entities when it is de-attached from the BS, and de-register MS from the network
when its context becomes unavailable in one of its serving function locations.
When the keep-alive mechanism is enabled the ASN-GW periodically polls other
ASN entities-of-interest (BSs) and waits for their responses. In case of no
keep-alive response, the ASN-GW shall make further actions, such as clearing the
applicable MS(s) context.
The ASN-GW builds a list of BS-of-interest which it must poll. The list shall be
dynamically updated; the ASN-GW tracks all BSID(s) in all MS(s) contexts it holds,
and dynamically updates the list of BSs-of-interest. When a new MS is attached to
a BS that does not exist in the list, it will be added it to the list. When the last
MS(s) with specific BSID makes network exit, the ASN-GW shall remove the BS
from the list if there is no other MS attached.
The ASN-GW periodically polls the BS(s) for keep-alive. The polling mechanism is
independent and unrelated for every BS-of-interest the ASN-GW polls.
The keep-alive mechanism uses configurable retry timer and retries counter.
Upon expiration of the retry timer, the ASN-GW resends the ASN Keep-Alive
request message. Upon expiration of the retries counter, the ASN-GW assumes
failure of the polled BS and clears the contexts of all MS(s) served by that BS.
In addition, the ASN-GW verifies that for each polled entity that the
“Last-Reset-Time” UTC value of poll N+1 is equal to the value of poll N. If the
“Last-Reset-Time” UTC value of poll N+1 is higher than the value of poll N, this
mean that the BS went through reset state during the interval between two
consecutive polls. In this case, the ASN-GW shall clear all MS(s) contexts, served
by that specific BS that are “older” than BS life after reset (through calculation of
difference between polled entity “Last-Reset-Time” received on poll N+1 and MS
network entry time stamp on ASNGW).