Selection menu to acces the graphs, organization of the graphs, title of the
graphs, association of a graph to a host that is not concerned by the
collect
============================================================================
==============
Hello,
For the questions see chapter (2) below
1) First I explain the context to better understand my problem
==========================================
I'm currently looking at the JF NMS from a user point of view ; for me
this point is very essential.
Effectively, I'm making an the SLA for the LAN.
In my study I have distinguished two kinds of metrics :
- The one that may interest the users and that don't need to be an
expert of the network (also I don't want to provide to the user the
architecture of the network, but just precise which user is connected to
which network device, and which main resources are concerned by the metrics
measurements)
The metrics I have seen are :
- reachability of the resources
- availability of their network device they are
directly connected to
- response time of the resources
- volume and throughput exchanged with the resources
- packet loss to reach the resources
- The other ones are expert metrics
They are such are
- network device cpu, memory
- traffic throughput on diffeent physical interfaces
- ...
So, my main idea is to make the tool that display graph useable by the
users which are not necessarily network experts
I want they be able to get very easily the graphs relative to the
network device they are connected to.
I haven't chosen yet the tool to do that. In my company currently we use
BMC Patrol Dashboard (but this is not enough flexible), Cacti (but this
doesn't send any alarm and doesn't allow making zoom and historic of the
graphs), then I have seen that JFF NMS allow to keep archive of graphs,
allow to make zoom, allow to send alarm, talk abour SLA ; that's why I'm
testing a little this tool.
1.1) Organization of the graphs
=====================
In Cacti for example, I would organize the graphs as descibed in the
menu below :
Building_1
Network_Device_A
SLA Metrics
- the 5 graphs above (1 per metric :
reachability of the resources, response time of the resources, ......)
Expert Metrics
- some metrics general to the device
(i.e. memory, cpu)
Interfaces
Interface_1
- some metrics relative to
the interface
Interface_2
- some metrics relative to
the interface
Network_Device_B
Building_2
I'm trying to organize the graphs like that in JFF NMS.
This in order to give focus on the SLA graphs dedicated to the user and make
him avoid distance by expert metrics.
When the user choose a network device is able to see immediately all the
graphs of the SLA metrics.
All the graphs will be displayed on the same page.
1.2) Title of the graphs
================
Also I want that the title of the graphs be meaningful for the users :
That is to say the must keep
- the name of the device they relate to (but this is not
necessarily the network device on which the measure is really done --> see
1.3)
- the name of the metrics (ie. Response time to group
resources A)
- the date and time of graphs + begin + end
For the same network device we may also have ("Response time to group
resources A" and "Response time to roup resources B"), that is to say
several graphs of metrics if there are different group of resources.
1.3) Why measure of the metric is not necessarily done the node
presented to the user ?
===========================================================
I want to hide the complexity of the network architecture to the user, and
just let him/her know on which device he/she is connected, and the main
resources he/she accesses, so I cannot present him/her all the interfaces of
the network devices, or additional devices or computer used to make measure
Effectively, the metrics won't necessarily be measured from the network
device the user is connected to.
For instance, lets assume the user is connected to a device DEV_USER, and is
accessing to ressources connected to device DEV_RESOURCE. I may put a
shadow SAA router (router called DEV_SAA_SHADOW) behind the resources, and
make on it some SAA probe to get response from a computer (COMPUTER_QUERIED)
that is always swithed on, and located near the user (behind DEV_USER for
instance).
USER_COMPUTER --------|
|------- RESOURCE
|------DEV_USER------------------DEV_RESOURCE------|
COMPUTER_QUERIED ---|
|------- DEV_SAA_SHADOW
So I will make the measure between the 2 following nodes in this direction
DEV_SAA_SHADOW ============> COMPUTER_QUERIED
Then I'll display on the graphs for the device DEV_USER, the response time
of in fact DEV_SAA_SHADOW --> COMPUTER_QUERIED.
Sure it will be an approximation, but it will be sufficient for the users.
1.4) My observations and problem I have met
==============================
From a user point of view, when we use JFF NMS, to display graphs we
have to choose (Host, Map,... for instance) then we have to choose a
particular "JFF interfaces" but this is not very meaningful for the users.
It is too much bound to the way that data are collected and the graphs are
displayed. It is a developer or JFF administrator configuration information
(which precises how data are retrieved) not a user information. Furthermore,
a single interfaces can lead to multiple graphs that make this still more
complex I think.
In the selection list on the left, I think that the results should be
one of those
DEVICE_A - Graph_name
DEVICE_A - Physical Interface_B - Graph_name
DEVICE_A - Service_name (ie. http) - Graph_name
DEVICE_A is not necessarily the DEVICE on which the measured is done ===> So
it is not necessarily the host_name from which data collection and graphs
are configured (that is to say not necessarily the host_name to which "JFF
NMS interface" belongs to).
In addition, the title of the graph itself should be the line selected.
2) My questions
============
How can we have a menu in JFF NMS, that allow the user to choose one of the
item below to display the graph ?
DEVICE_A - Graph_name
DEVICE_A - Physical Interface_B - Graph_name
DEVICE_A - Service_name (ie. http) - Graph_name
(DEVICE_A is not necessarily the DEVICE on which the measured is
done )
How in JFF NMS can I have graph titles that exactly correspond to the item
selected above ?
How can I make measure/collect on a DEVICE_A and associate the results to a
DEVICE_B ?
That is to say display the results whern I select DEVICE_B ?
Nevertheles, I thanks a lot the developers that have worked on this
application. I know that tasks and integrations of the user expections are
numerous and not necessarily easy to take into account. Already I find that
a very good job have done on this applications compared to the others
applications around the RDD tool.
Thanks a lot for any answer
Didier
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
jffnms-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jffnms-users
