CIO and Perfcastination

Ask any CIO, how much time they / their team spend on improving the performance of the application deployed. Lowest answer you can get is 25% of their total time.

Ask any CIO, what are the top 5 priority items on his / her list? I be surprised if one of them is not applications performance.

Ask any CIO’s if they have any existing request for hardware upgrade on his / her list to solve performance issues? Answer will be yes.

Over a period of time, based on my experience I have found that most applications are never designed for performance. Development team’s focus is functionality. In quest for packing more functionality they loose sight of performance requirements. Most cases they need to deploy the solution to start the business process. Application deployment happens in most cases without performance testing. Few weeks into the operations, performance issues start showing up. When asked the common answer to these performance issues will be

A. Application is running on low end server
B. Application is sharing the box with other applications
C. Backend database is not capable of handing the load
D. Disk speed is slow
E. Not enough RAM on the server
F. Network link between the application and / or to the internet is slow

In most cases – the panacea prescribed is “Upgrade and your performance issue is solved”

This is nothing but what I called “Perfcastination” by software development team. CIO’s team goes hunting for budget. Most often hardware / application are recently moved into production. Getting budget approved is a nightmare for CIO. Two common questions are

Why did you not right size it just few months back?
Do you guarantee that this upgrade will solve the issue?


Perfcastination is perfect tool used by development team to park the problem. Most often when performance issues happen, it’s either CPU is fully utilized, high disk I/O, longer SQL execution time etc. Easiest thing is to park the problem on the upgrade that gives them couple of weeks to months to get it done. While some smart teams use the time to look at the intrinsic issue and resolve it, however most bank on the upgrade. They truly believe in what they preach!!!

Post upgrade, most often than not, CIO is again back to the desk with performance problems.
I leave you readers to read this and let me know your thoughts.
My next post on the topic shall be - “101 on managing Perfcastination for CIO!”

Asiatic Lion

Sunday - 12th Oct was a day with little activity.
While browsing the net around to see how much information is available about my native place Una - Gujarat. I landed at two good sites giving information about Asiatic Lion and Gir forest.

Gir Forest Guide - http://girforestguide.blogspot.com/

All about Asiatic Lion - http://girasiaticlion.blogspot.com/

I had the opportunity to see Lion's in Gir forest couple of times. Grandeur of 'Vanraj" as called in Gujarati cannot be explained in words here...It can best be seen...

As it stands today there are close to 350+ Asiatic Lions in the world today. If you spend couple of day's in the forest, you are bound to see them multiple times with their prey or at water ponds...

Airtel eBPP application reuse- more findings.....

One of my friend did not see the Airtel+Vodafone logo, I saw on Airtel's eBPP site. We both were true, the magic was this script, based on the system time zone, it will show

https://ebpp.airtelworld.com/wps/airtel/images/Airtel_Jersey_logo.gif

or

https://ebpp.airtelworld.com/wps/airtel/images/airtel_logo.gif

Word Jersey in the logo's file name made me think that Airtel may have some tie up with Vodafone in Jersey, google my best buddy on the net, validated my gut - http://www.vodafone.com/start/media_relations/news/group_press_releases/2007/jersey_airtel_brings.html.

So if you are in timezone on or around GMT (which was my system setting) it would show Airtel+Vodafone logo and for rest it would show Airtel logo.

Now this seems like sensible reuse of code, except for the generalization that all those who use GMT timezone are from Jersey and rest are in India :)

Heights of reuse

I recently became a corporate customer of Airtel. To view the bill details, from Airtel’s site, I was redirected to https://ebpp.airtelworld.com/ to my surprise I saw a logo of Vodafone too on the website.




This is a classic case of reuse if Vodafone customers too access this same URL to view their billing information. I vaguely remember that IBM does IT support for both Airtel and Vodafone. This could be a classic case of reuse 

Hope the branding guys at Airtel and Vodafone are checking this out………or this is a part of share the infrastructure project.....

job board and their rankings

Posted this note on linkedin to a question from one of the member. Sharing this on my blog too.

Each job board has unique characteristics of candidates it attracts. I have only used India specific job boards. So the thoughts that I am sharing are India specific. However some of them would apply in general. Most of the boards in India complete on 1) number of resumes in the database and 2) number of visitors to the site. I have found that with growing count of resumes on the job board, we have to dig out a lot to get to the right candidate. Numbers of visitors are dependent on number of job postings. As mentioned by other members on this forum, recruiters rule most of the job boards. They keep posting same jobs again and again.

So really to me, job board ranking based on visitors do not make sense at all. While selecting job board for recruitment I have used the following approach.

Got multiple job boards to give a trial access, Looked at various skill sets, briefly spoke to candidates from those job boards. Based on the evaluation of ease of use, kind / quality of resumes and type of candidates that have submitted the resume I selected the job board.

I feel that job board ranking in my opinion do not make any sense as parameters used for ranking and parameters that determine the quality of candidates that we want do not have any correlation

Mobile service with limit on voice and SMS - will it work in India?

Today AT & T announced ( that theyhttp://money.cnn.com/2007/09/04/news/companies/bc.at.t.parentalcontrols.ap/index.htm?cnn=yes ) will have parental control :) on cellphone usage by their Kids. Of course this feature comes at monthly recurring price. At the face value looks like a good option or rather an insurance for not getting higher monthly bill. Question that comes to my mind is will it work in India. I think it will not.

1. The mobile operators would not like to invest in the operational software needed to limit the usage.
2. With the prepaid cards as the preferred option in India, the limit is automatically set.

However only time can prove me right......

Training BBM students at Surana college

What else can be a better way to start a week than spending time with students. Today I was at Surana college, teaching final year BBM students 'Interview skills'. Was good fun and they had real good question. I saw them no different from the fresher's who I interview at work. Some had their head firm and while others who need to do some fixing. Overall a great morning after locking my car with the key's inside. Rahul thanks to your team it went all smooth!.

Roaming cheaper than local calls - It only happens in India

Over the weekend I was thinking, the type of choices consumers face in this maze of telecom price plan.

I found that calling landline within the circle (for a plan) is costly, when compared to the same user going out of circle and making a local landline call in that circle using roaming service.

To give an example, SPICE Telecom in karnataka has a Plan called 'Spice Uth Extra' (http://www.spicetele.com/karnataka/PD_uthextraTariff.php). It has local landline calling rate of Rs. 1.69 per minute. However the same subscriber / customer can go to TamilNadu / Goa / Delhi / anywhere outside Karnataka and call land lines in that circle at Rs. 1.40 per minute (http://www.spicetele.com/karnataka/national_prepaid_roaming.php).

Does this not sound funny?

In the case of local calling within the circle at the max only two telco's are involved viz. mobile provider and the called landline provider.

Where as in the case of roaming, it could involve max of three telco's viz. roaming service provider, customer's service provider and called telco. With three people to share the revenue, should that not be costly?

I am sure that Telco's may have their own data mining and marketing strategist's that plan out such tariff plans, however for the poor consumers its a lot of effort to select a plan that fits him / her