| 2 |
|
=================================================== |
| 3 |
|
|
| 4 |
|
ajm4, 30/10/2000 |
| 5 |
+ |
tdb1, 30/10/2000 |
| 6 |
|
|
| 7 |
|
System Component Startup |
| 8 |
< |
------------------------ |
| 8 |
> |
************************ |
| 9 |
|
|
| 10 |
|
CORE |
| 11 |
< |
**** |
| 11 |
> |
---- |
| 12 |
|
|
| 13 |
|
|
| 14 |
|
Client Interface |
| 15 |
< |
**************** |
| 15 |
> |
---------------- |
| 16 |
|
|
| 17 |
|
|
| 18 |
|
Filter |
| 19 |
< |
****** |
| 19 |
> |
------ |
| 20 |
> |
The filter is broken down into three main subcomponents. |
| 21 |
|
|
| 22 |
+ |
- Filter Manager |
| 23 |
+ |
The Filter Manager is responsible for managing which |
| 24 |
+ |
filters are used by the hosts. The Filter Manager is |
| 25 |
+ |
available at a "well known" location which is pre- |
| 26 |
+ |
programmed into the hosts. The Filter Manager is |
| 27 |
+ |
responsible for creating and managing the other |
| 28 |
+ |
components of the filter system. |
| 29 |
+ |
|
| 30 |
+ |
- Main Filter |
| 31 |
+ |
The Main Filter is the single point that links back |
| 32 |
+ |
into the CORE of the system. It will connect to the |
| 33 |
+ |
DBI and the CLI to deliver data. |
| 34 |
+ |
|
| 35 |
+ |
- Filters |
| 36 |
+ |
There can be multipler Filters, and these are the |
| 37 |
+ |
"front line" to the hosts. They all link back to the |
| 38 |
+ |
Main Filter to send data into the system. It is |
| 39 |
+ |
possible to run these Filters on any machine, allowing |
| 40 |
+ |
management of data flow. |
| 41 |
+ |
|
| 42 |
+ |
At startup a Filter Manager object is activated at the "well |
| 43 |
+ |
known" location (probably a given machine name at a |
| 44 |
+ |
predefined port). The Filter Manager will create an instance |
| 45 |
+ |
of the Main Filter, and any Filters under it's control. |
| 46 |
+ |
Through some mechanism the other Filters, elsewhere on the |
| 47 |
+ |
network, will register with the Filter Manager. The |
| 48 |
+ |
Filter Manager will need to tell each Filter the location |
| 49 |
+ |
of the Main Filter upon registering. The Filter Manager will |
| 50 |
+ |
then be in a position to receive connections from hosts and |
| 51 |
+ |
pass them off to Filters. |
| 52 |
+ |
|
| 53 |
|
System Running State |
| 54 |
< |
-------------------- |
| 54 |
> |
******************** |
| 55 |
|
|
| 56 |
|
CORE |
| 57 |
< |
**** |
| 57 |
> |
---- |
| 58 |
|
|
| 59 |
|
|
| 60 |
|
Client Interface |
| 61 |
< |
**************** |
| 61 |
> |
---------------- |
| 62 |
|
|
| 63 |
|
|
| 64 |
|
Filter |
| 65 |
< |
****** |
| 65 |
> |
------ |
| 66 |
> |
When a host first loads up it knows where to locate the |
| 67 |
> |
Filter Manager because it's located at a "well known" |
| 68 |
> |
location. The host will fire up a TCP connection to the |
| 69 |
> |
Filter Manager to announce itself. The Filter Manager will |
| 70 |
> |
use some method (logically) to allocate a Filter to the |
| 71 |
> |
host. The Filter Manager should base this decision on |
| 72 |
> |
various factors, such as the load on the selection of |
| 73 |
> |
filters, and possibly the location in relation to the host. |
| 74 |
> |
The host will then be directed to this Filter for all |
| 75 |
> |
further communications. |
| 76 |
|
|
| 77 |
+ |
As the system runs the host will send data with (maybe) UDP |
| 78 |
+ |
to the Filter (that it's been allocated to). This choice has |
| 79 |
+ |
been made because it puts less onus on the host to make the |
| 80 |
+ |
connection, rather the data is just sent out. However, to |
| 81 |
+ |
ensure that the data isn't just disappearing into the depths |
| 82 |
+ |
of the network a periodic heartbeat will occur (at a |
| 83 |
+ |
predefined interval) over TCP to the Filter. This heartbeat |
| 84 |
+ |
can be used as a form of two-way communication, ensuring |
| 85 |
+ |
that everything is ok, and if required, to send any |
| 86 |
+ |
information back to the host. This heartbeat must occur |
| 87 |
+ |
otherwise the server may infer the host has died. |
| 88 |
|
|
| 89 |
+ |
This could link in to alerting. An amber alert could be |
| 90 |
+ |
initiated for a host if the server stops receiving UDP |
| 91 |
+ |
packets, but an red alert be raised if the heartbeat doesn't |
| 92 |
+ |
occur. |
| 93 |
|
|
| 94 |
+ |
If, for some reason, the Filter were to disappear the host |
| 95 |
+ |
should fall back on it's initial discovering mechanism - ie. |
| 96 |
+ |
contacting the Filter Manager at it's "well known" location. |
| 97 |
+ |
The host should report that it's lost it's Filter (so the |
| 98 |
+ |
Filter Manager can investigate and remove from it's list of |
| 99 |
+ |
Filters), and then the Filter Manager will reassign a new |
| 100 |
+ |
Filter to the host. Communication can then continue. |
| 101 |
+ |
|
| 102 |
+ |
The idea of plugins to the Filters has been introduced. |
| 103 |
+ |
These plugins will implement a predefined plugin interface, |
| 104 |
+ |
and can be chained together at the Filter. Using the |
| 105 |
+ |
interface we can easily add future plugins that can do |
| 106 |
+ |
anything from parsing new data formats, to implementing |
| 107 |
+ |
encryption algorithms. The Filter will pass incoming data to |
| 108 |
+ |
each plugin in turn that it has available, and then finally |
| 109 |
+ |
pass the data on to the Main Filter. The Filter need not |
| 110 |
+ |
have any real knowledge about the content of the data. |
