Home
> Technology > Transaction
Path Analysis
Transaction Path Analysis
The MetiLinx Sensor can do more than just monitor
the available resources on a single server. It can
leverage this information across the entire path of
a transaction. Let's take the example of a web tier,
business application tier and database tier. When
the MetiLinx software is first installed and configured,
the administrator identifies pools of similar servers;
then identifies the paths those transactions take
between pools.
The MetiLinx Sensor calculates the Local Node Value
for the amount of free resources in each server.
Server |
LNV |
Web_1 |
44 |
Web_2 |
41 |
Web_3 |
43 |
|
|
Server |
LNV |
App_1 |
57 |
App_2 |
58 |
App_3 |
55 |
|
|
Server |
LNV |
DB_1 |
66 |
DB_2 |
61 |
|
|
|
After calculating the amount of free resources in
each server, it sends that Local Node Value "upstream",
while at the same time recording the network latency
encountered while sending those values. By combining
the resource and latency values the MetiLinx Sensor
calculates a "Composite Node Value" (or CNV) that
represents the amount of free resources for the transaction
path.
Using the example, we get the following calculations:
DB_1 CNV = DB_1 LNV because there is nothing "downstream"
DB_2 CNV = DB_2 LNV because there is nothing "downstream"
App_1 CNV = a calculation from the LNV of App_1, the CNV of DB_1 & DB_2,
and the network latency between App_1 and DB_1 & DB_2.
This Composite Node Value represents the amount
of free resources available for transactions on App_1,
plus the network latency and available resources
on servers downstream. This provides a much more
complete picture of the transactional system than
looking at a single server's utilization value. It
also provides a depth of information that is quite
impossible for an administrator to gain from looking
at CPU, memory and disk values for individual servers.
Server |
LNV |
CNV |
Web_1 |
44 |
55 |
Web_2 |
41 |
53 |
Web_3 |
43 |
54 |
|
|
Server |
LNV |
CNV |
App_1 |
57 |
62 |
App_2 |
58 |
63 |
App_3 |
55 |
60 |
|
|
Server |
LNV |
CNV |
DB_1 |
66 |
66 |
DB_2 |
61 |
61 |
|
|
|
|
By doing such an in-depth and cross tier analysis every second, MetiLinx is
able to find the overall transactional path utilization and report it as
a single value.
To complete the example, Web_1 currently has 44%
of its resources available locally, but the transaction
path starting at Web_1 has 55% of its resources available
(i.e. Web_1, plus business application tier plus
database tier plus network latency).
|
