With Open Source One needs to be Generous -OSI days Bangalore 2014

Attended the OSI days Bangalore. One of the places where it makes sense to host a conference where you can find a large numbers of people on any technology under the sun. Open source unifies the birds working on different tech stacks .NET , JAVA , ROR, PHP, LAMP,MEAN folks [ how mean or cheap can you get 🙂 with open source ] etc. Scripting all shades and Mobile App developers worth their app.

It was organized by EFY the magazine company primarily with other Open source vendors / sponsors with their stalls.Microsoft had a big stall along with Oracle and HP all out there hustling and trading their wares enticing developers to get hooked on their bandwagon. Microsoft had a dev camp where a set of exercises were loaded onto PCs which was open to developers to type , program the exercises and get a real world feel of how quick or easy it is to get your app moment. There were applications ranging from test programs to play around with Mongo DB, windows HTML mobile applications. Nice way to get developers to at least take a peek at your offerings.

And there were goodies like T-shirts, portable chargers and like to make it worth somebody’s time to try them.

It is not about who makes the box anymore.

HP Helium was out their doing their social marketing around the Helium Open Stack cloud. They have a strong private cloud offering with Open Stack as the standard and a lot of evangelizing around that was there to see.Tweet to get early bird prizes,quizzes,selfies randomly picked and prizes ranged from HP tablets to laptops in some cases. Over all it was a big change to see traditional companies playing the social game well , getting public attention on social media. With game changing and disruptive technologies all around HP has thrown their hat in the cloud arena with the Helium offering. As it is now “It is no longer about who makes the box”.The infrastructure game has changed forever with cloud adoption. Nobody wants the CAPEX with an expensive box in their premises unless it has a strong rational these days.The only way HP can push their Hardware capability into the markets is for people to go the private cloud way and then push their servers where it has a business use case. Overall a lot of action happening on the front with HP having partnered with Cloud foundry to make it happen among other key cloud initiatives Good to see large companies stepping up their game or risk being outpaced. Microsoft adopting the cloud game with Office 365 was spot on and help them stay in the game. Sure HP played their social media strategy well here and more so get developers to talk about you and the market will follow suit.

Notable misses from an Open Source conference were Redhat,Google,Python community but nevertheless it was a good representation overall. May be it is all that you can squeeze into a two day program with back to back parallel sessions.

Some random takeaways

1. Companies of all sizes and shapes are adopting open source. Some large companies now have an open source practice, Wipro had a stall there.

2. People choose the technology stack based on what they are comfortable while building products and not necessarily the best out there for a given problem. It is difficult for one to know what is good with rapid technology changes.

3. The shelf life of any product built with Open Stack is only a few months to a year. Then on they either get upgraded or move to something better.

4.Mostly backend business logic is still written in good old C,C++,Python,Java,COBOL and people do not tamper with them although new technology intrusion is always enticing.

5. GoLang preferred against Python where concurrency was an issue.Although Python measured up against GoLang with no CPU latency being there.

6. A lot of product ideas being thrown around and some folks detailed what their journey was like. Good comment was when someone said you adopted SMART technologies to get here, the response was it took us several years to get SMART. So it was never a overnight thing.

7.Stackoverflow and user forums for the Open Source technologies apart from Google GOD solved people steer clear of roadblocks and bottlenecks during their open source journey.

8. Although getting people to adopt open source is hard with no support but once the team adopts to the stack it works great and an unmatched ROI. Else your TCO is eaten away by the hardware/software vendors whom you are depended on. One mention here is a tool called sendy which costs around 60$ but once configured and setup you can send mail blasts to your email list for a lifetime with that initial cost.

9. You need to be generous to allow people to use your stuff for free and this promotes your to leverage on the collective strengths of an intelligent community whom you can bank on for updates , fixes , issues etc.

10. Leveraging Open Source is a two way street it always works with both the parties being benefited Once you have an enhancement that serves the larger community you contribute back into the pool. Opportunity here is a two way street.

People pay for the shiny stuff large user base gets you marketing

You need to be generous to allow a large percentage of people to use and benefit from your goodies. Evernote , all the cloud storage option providers ( gmail, one drive, dropbox) all of these fall into this category. They get their revenues based on the premium few users and good ad strategy for them to allow people to use their stuff. This is more popular as the freemium model, keep a large part of it free but for the shiny wares on your stack charge a premium. You need to give to get back. It works the other way also as you can almost crowd source efforts on your development if the open source takes off. You do not have to employ the services of people by employing people to take care of enhancements , future releases, bug fixes. This is a good advertising strategy for people who do not have ad revenues. Keep your product free , get people to comment on it crib and better still have them validate it for you free and then decide which of the features of the product you want people to pay for it. Quite disruptive coming to think of it , if its well executed.

JUDCON 2014 Bangalore – The Future is Open

Yes the future is open and that is what open source tries to create and emphasize. The attendance at the conference was large and many developers JBOSS / java folks were present. Developers poured out in large numbers and the air raved and ranted about goodies from the JBOSS stack and how RAD ( Rapid Application Development) can be had by using these cool tools from Redhat. This was my first to a redhat conference although have interacted with them on multiple occasions especially during a porting assignment in one of my previous avatars.

This was well attended from developers , architects and from all interested in what is latest on the JBOSS stable and how best they can integrate it into their products or solutions.

Attended select sessions from speakers and here is some brief summary of what they looked like . It is important for technology decision makers and people connected at all levels to the task of producing meaningful software to be abreast of what is latest from a popular open source vendor like Redhat.

OpenShift a PAAS solution for JBOSS developers : This is meant to have a all in one software plugins on your development environment ( maven, dependencies, libraries,deployment scripts,build scripts ) all bundled on a platform which is supposed to cut down on the dirty element in the process of churning software which is that many developers put in many many hours at creating software and lost countless hours creating different setups to be able to pull software and build them effectively. There needs to be a lot done in the PAAS area and this space will have lots of interesting things over a period of time. Eventually where thing including our local eclipses and netbeans will be on the cloud.

JBOSS is being rechristened as wildfly for the application server community edition and thus differentiates the Enterprise edition. JBOSS is no longer the application server company it has many many things under its hood and therefore justifies the change in name.

Pair programmed with Bharath of bharathwrites.in. on using widlfy and how to create an application which was organized and supervised and guided from Arun Gupta and Andrew of redhat. The wifi had its share of issues and folks running around with USBs to work around the problem. Annotation will rule the world as Java 7 has more annotations and more of it. Finally you could circumvent a lot of coding using annotations. The future would be annotations the way it goes for java and probably for other JVM languages.

Some random key take aways :

1. Websockets removes the overhead of HTTP / TCP and is preferred when you need to heavy connection based message passing. This can greatly improve performance and remove the clutter from the traffic especially the header and footer going back and forth for other protocols.

2. Batch Processing is a new feature of Java EE 7 apart from websockets, which gels with what JBOSS has to provide.

3. CDI Extensions ( dependency injection ) and add on features on how you can create features needing coding using annotations. Getting used to the latest ones are an issue but soon they can do a lot with no coding. The future is also open and annotated should be stated.

4. Fine tuning case study of a telecom major in India tuned all their way across the application stack across all tiers mod_jk , native APR , web subsystems , H/W accelerators , SSL accelerators, NUMA architectures , increased / created optimum memory page sizes on the OS, used sticky sessions and replication where possible and increased performance. This reminded my own experience where every performance increase was considered a mile crossed and when every inch , every ms every byte mattered. All of this can be useful when hard pressed for increased performance. Had interacted with clients who had spent a fortune on a machine and their only concern was did not want to move to a new machine and wanted to squeeze every pie out of the existing machine.

5. A nice case study from Wipro on how messaging solved many real time problem scenarios and how the Smart Santander a project in Spain uses much of this technology with sensors and devices to solve a lot of governance problems. Apache Camel and Active MQ apart from Kafka which was used as the ESB here was used to solve these problems.

6. Attended a nice talk on Infinispan on Inmemory database. At the outset the new statement was “Memory is the new disk and disk is the new tape”. Has a lot of use with transactional and non transactional scenarios with support for the No SQL and regualar RDBMs databases. This greatly improves performance and has a lot of settings on storing data close to where it is used and fetched using a hashing technique. This gives a uniform experience to users who could log in from any of the nodes in the system and almost be guaranteed a regular fetch time instead of varying fetch times. In memory would be the way for data fetches including transactional data which does not change too much.

People walked with hoodies and cool T-shirts apart from some lucky draw events from Intel on whose machines redhat linux largely gets run from most organizations. The future is open and this will a game changer on how to package your good small and service based. As some futurists predict there will be only three movements in future publishing , content and services based on that content. We are in interesting times where products/solutions designed to deliver value in portions and do not quote an hefty price upfront. This is the new selling mantra. And this is here to stay. Everything will be a service hardware/software/infrastructure/everything you model based on this principle can be attempted. Perhaps in future there could be a central air conditioner in a flat which could bill inmates on pay as you use and avoid people owing one and having to pay the total upfront cost when all they use it in places like Bangalore during the hot months.

About Us

All organisations have an innate intelligence which enabled them to become an organisation in the first place. Most often organisations who reach some height do not get their steps right in terms of moving in the turns often failing to tap their intelligence. Intelligence in organisations and enterprises is hidden and disconnected and at times do not augment the need of the hour. The fallout would be enterprises fail to step up to the current challenges which once enabled them to stay ahead of the curve no longer seem to work. We often see what works to maintain a company of 10 is not enough when the numbers go up. We have all sorts of challenges right from knowledge management, isolated knowledge pockets, non modular legacy, all pieces glued together,scaling up the build , hit the right numbers on performance , get the architecture right , do we have people with the right skills and knowledge and many more.

All of these need to tap the organisational intelligence at various points and dovetail them to align with the business and eventually technology needs to enable this to provide value to the business, stakeholders and customers.