Efficient Solutions

http://www.computer.org/portal/web/computingnow/onarchitecture

Is anyone able to find it?

http://www3.opengroup.org/events/our-webinars

Very interesting link:

http://docs.google.com/viewer?a=v&q=cache:4mM_K5Kt-tUJ:www1.eere.energy.gov/femp/pdfs/eedatacenterbestpractices.pdf+top+energy+efficient+servers&hl=en&gl=us&pid=bl&srcid=ADGEEShPlfLpJG9n3PucDkRH0_qobNWD0slI37SnKxaQjsDUdG6b6KdskMLxHAnBgXa35lzjSH_NvC68T_GY6zDlLtdNydlESOiaFHKERdCrjME44yjBnpKbVUdk1ncDy-PnY-dm1VnC&sig=AHIEtbRSAZ_D3mE8dvsiosw7pdlnLDbUQg

TBSM is one of industry’s most looked upon product when it comes to service visualization, service level management, SQM and lastly BSM solution. I have worked on TBSM since early RAD 3.0 days when only a handful of folks really worked on it, and a ton of features never worked (for example the mapbuilder.sh in RAD 3.0).  Times evolved and so has the product, built on features like service templates, service instances, auto-population rules, ESDA and some decent visualization templates. So much about the history, where are we heading! What is the future, what will be the place of TBSM in future, what does TBSM need to do to ‘lead’ not just exist, these are some of the questions, I will be putting my 2 cents into.

Lets bring some context here, the concept of ‘service’ gives most executives a kick and selling TBSM is not really rocket science, when you can view the business scorecard from technology, operations and financial context in a solution, the solution has to sell. But that is just the very first phase, it is the post sales, execution that is the key. We need to put some goals that define the finish line, so what are they – Is it to be able to leverage the solution to it fullest or is it systematic step by step setup of its BSM context and visualization of services, financials and efficiencies? This is the area which is the burning question and all said and done still an area that needs to be addressed together by many stakeholders. So how to make this all happen – is it even possible to have a common TBSM deployment framework which would be generic enough to fit any bill and yet not broad enough to overcommit and under deliver – my belief 100% YES!

To summarize, my experience with TBSM has always been that it requires a ton of specializations and generalizations, understanding the contexts, abstraction layers, encapsulation of information, but still all environments have a common theme. Unfortunately noone has really built a common deployment process framework (1,2,3 step-by-step deploying TBSM). Unless, the users  and TBSM services professionals come together and build this framework, the quality of delivery will never be consistent and we will keep bumping into dissatisfied clients. Shout out  for TBSM deployment framework, not from the PDU teams but from the services professionals, users and process analysts who deal with reality on a day to day basis.

Of late, I have been looking at Datacenter automations and impact of DCA on costing for the datacenter.

Various solutions exist, to name a few Emerson Aperture, NLyte and Rackwise – All of which bring important things to the table. But what are the big 4 doing in this space? Do they have any solid solutions out there? Or they are just living with the situation?

Lack of investments in RFID, rack monitoring software has really taken the cost of manually managing datacenter’s higher. The bottom line is datacenter is a piece you cannot just outsource because it is the fundamental part of unique value chain for every company.

The only proposed solutions are:

1. Virtulization mgmt
2. Workflow management – Cable mgmt, Physical and logical asset management
3. Spare management
4. Space, Power, Cooling mgmt
5. On-boarding, offloading management
6. For cloud – customer centric solution management

Unfortunately the quest for finding the right solution continues, but I am not giving up.

Some product/consulting companies charge upto 25K USD for integration of FM-FM/FM-PM products. One has to be careful of such offerings because not only they have a one time cost, but also they come with a continual license fee for the gateway. BAD!! So let me save you some money by generalizing this process by an example of integrating two highly used NMS solutions – Tivoli Netcool [from IBM] and NAGIOS [Open source offering]. Integration from Nagios to Netcool is simple [not sure why people pay tones of money for this] and can be done in couple different ways:

Overview

1. Asynchronous uni-directional data flow [from Nagios SBI to Netcool] : In this method of integration, Netcool shall receive events  as forwarded, but shall not acknowledge the event back in Nagios. This is useful when Nagios is not used by operators for RT monitoring.
2. Synchronous bi-directional data flow: An event in Nagios will flow to Netcool and will be confirmed back in Nagios as recieved by Netcool. On every update on the event [such as journal entry, acknowledgements] the event in Netcool, status shall be updated in Nagios.

Either options work based on the business/solution requirements. So without further ado:

Implementation:

1. Asynchronous uni-directional data flow [from Nagios SBI to Netcool]

To understand the implementation, I shall divide the steps as southbound implementation and northbound implementation. Southbound implementation refers to the changes/configuration on Nagios end, and Northbound implementation refers to updates in Netcool.

Southbound updates [On Nagios];

a) Create a script to send tcp socket messages or snmp traps or direct JDBC insert to NBI.

You can use snmptrap command for writing the script, if you are not a SNMP guy you can use a simple script to do socket message communication/JDBC inserts into Objectserver. Test this script.

sample snmp script:

Send trap

# Arguments:

# $1 = Management Station

# $2 = Community String

# $3 = host_name

# $4 = service_description (Description of the service)

# $5 = return_code (An integer that determines the state

# of the service check, 0=OK, 1=WARNING, 2=CRITICAL,

# 3=UNKNOWN).

# $6 = plugin_output (A text string that should be used

# as the plugin output for the service check)

#

# Sample

# /usr/bin/snmptrap -v 2c -c $2 $1 ” NAGIOS-NOTIFY-MIB::nSvcEvent nSvcHostname s “$3″ nSvcDesc s “$4″ nSvcStateID i $5 nSvcOutput s “$6″

b) Define a global event handler in Nagios: Global event handler will help execute the script on every state change on Nagios instance and will communicate, failure and seizure of the problem. How to configure GEH: http://nagios.sourceforge.net/docs/2_0/eventhandlers.html

Northbound updates [On Netcool]

If SNMP:

a) Download the Nagios MIB and compile with MIB2Rules

http://sourceforge.net/projects/nagiosplug/files/nagiosmib/

b) Update the rules file and include it  in mttrapd main ruleset

If socket:

a) Update the socket probe to parse message based on delimiters

b) Ensure all mandatory objectsesrver fields are accounted for

If JDBC:

a) Ensure all mandatory objectsesrver fields are accounted for

b) **CAUTION** Watch the objectserver profiler for IDUC consumption, as this is not so much of a conventional approach

DID YOU CATCH THE HEADFAKE?

Nagios an Netcool were just examples, you can integrate most FM-FM/FM-PM solutions using the aforementioned procedure, you just need to know the NBI data model, SBI data model, right triggers on the SBI system and right listner on NBI system. Made your life easy, din’t I? So start saving your company some money now!!

In the next post, I will talk about method 2 {bidirectional data flow}. Keep visiting!!