Off the Shelf: The Retail & CPG Blog

EAI (enterprise application integration) is an interesting leg in today's IT landscape. I remember the days when started my carrier 13 years back when we all wandered between mainframes, AS400, database and new age object oriented java. With the passage of time I focused my area to be more specialized as the demand of market was to identify from the lot.

Java specialization was becoming readily available as most of the application started becoming loosely coupled. With this said tier based application was the faces of IT in late 90's and start of 2000. With this distributed architecture EAI found its place. I decided to focus myself in EAI technology after putting my fingers in AS400, Java and database technology.

Since then I have been following EAI and its progression. EAI vendors have increased over the period and the stack based companies like SAP, IBM, and Oracle have come up with their own middleware. The top notch players were Tibco, BEA, Seebeyond, webMenthods, Vitria etc. Along with the same group IBM had is Websphere in the market. Enterprise offering company joined in the game late and most of them tried acquiring the independent vendors.

Middleware originally structured itself by hub and spoke or bus modal. With the new age offering SOA, shock wave transmitted within the EAI space.  The maturity of SOA was highly dependent on the EAI. The reason being was it actually in true sense de-coupled the application. It was already playing an orchestration role. The only thing needed was the face lift of middleware to SOA suite. It was not easy as said to change the face. It came with its own challenges.

Traditional EAI was even though decoupled the system but the internal offering was vendor dependent. Each vendor interpreted the transmission in its own ways. This led the dependency heavily on the tool. The challenges started visible when transformation business started within the organization. Industry started means of saving cost, improving efficiency and reducing redundancies by consolidating into ERP packaged implementation. During this this time the need of standardization was realized. Where ever the ERP packaged solution was implemented its native EAI came by with cost saving measure. The architecture community was put in spot once again to re-define the co-existence of dual middleware strategy. Lot of company decided to migrate to single EAI this became the revenue for the consulting companies. Even though the data transmission was standardize to an extent but the transformation, orchestration was native to the EAI vendor.

SOA guiding principle made an impact on the middleware strategy. The aspect of re-usability, standardization and discoverability was re-invented for middleware. There was a push within the community to standardize the transformation and orchestration code. Evolution of Standard Integrated Development Environment (IDE) came in place. Eclipse (IDE) for example widely accepted with major IT vendors. Most of the product vendors shipped their software with standard IDE environment with product specific wrappers. This internally became standardization of code.

The future of EAI is now redefined as SOA middleware. The development process is standardizing within the SOA Middleware. The output code (like BPEL) can be ported within any SOA Middleware. The discoverability, standardization and re-usability indeed changed the course of traditional middleware. The openness now largely welcomed by the architect group. Business is now getting more agreed upon technology changes.

Written by and submitted on behalf of Mukundan Iyengar

Posted by Prasanna Ramaswamy at 4:53 AM | Permalink | Comments (0) | TrackBacks (0)

When I started my IT career in Infosys 11 years back and was allocated to the mainframe training batch and ultimately to a mainframe development project, I felt like an outcast banished from human civilization. For the love of God, why should I (of course along with bunch of others) have to learn and work on some technology older than my predecessors? To add to the grief, the entire world (well...almost) was holding up banners stating "mainframes are like dinosaurs...the end is near" or "MVS and its dumb terminals - rest in peace". Rest of my batch mates, who gained (not earned) the glory of working with 'open' systems used to look down at us with sympathy. Poor souls have to work on a technology which will be nothing but folklore in a few years - so they thought.

How wrong were they? And anyone else who thought that the age of the majestic dinosaur was gone. Mainframes are well and truly here, kicking and alive and will be for years, if not centuries, to come. Open systems and distributed platforms have been around for a long time, but have any of them gained the same level of trust and confidence of its masters as the mainframe? I doubt it. Just look at the big corporations who have to process billions of transactions with tight service levels to run their business. Majority of the banks, financial institutions, big retailers, health care and insurance companies still run their core business off the mainframes. Nothing has stood up and proven capable of taking over from these time tested behemoths.   

Apart from being the workhorse that it is known to be, the mainframe is highly regarded in terms of reliability, availability, serviceability, scalability, performance and security. This explains why the above mentioned industries rely so heavily on mainframes even today. What else can handle such scale of transaction processing per second? What else can efficiently support thousands of concurrent users and automated processes? What else can manage and process terabytes of data between the different layers of a complex application? What else can be relied upon to be available for a long (really long) period of time without having to mess with hung transactions, system reboots, software patches and OS crashes? Well....nothing.

Most likely, every one of you reading this article might have interacted with a mainframe at some point - when you did an ATM transaction, accessed an ecommerce site, paid your insurance premiums or did your loan calculation online. People crack jokes about dumb terminals and green screens. Well, green screens are long gone and seldom do these folks realize that green screens were the initial CRT monitors used to access a mainframe, which have been obsolete for decades. Nowadays, those have been replaced by LCD monitors and touch screens. About dumb terminal, how different is that from a PC connected to some application hosted on a virtual environment, a.k.a. cloud? Now that's state-of-the-art, isn't it? But, a dumb terminal connected to a mainframe server situated at the other end of the globe isn't? What a paradox! People...it's the same philosophy. Just that virtualization and centralization have been around for decades since IBM released System 360 back in the 60s. Old wine, new bottle.

Of course, mainframe had its drawbacks, with the foremost among them being cost. Almost all other factors have been nullified or reduced with technological advances and improvements. Additionally, mainframe boxes have become smaller and much more energy efficient, thereby ensuring efficient data center utilization. Mainframes have even become less expensive with these improvements, but remember that cost is relative. If you consider just the cost of hosting and managing a mainframe environment, that might be prohibitive compared to some of the other platforms in the same landscape. However, when you divide the cost by the number of transactions processed by the systems for the same time period, there will be very few platforms which can beat the mainframe's cost per transaction. A similar exercise done on a customer's mainframe system gave the cost per transaction as less than 4 cents/transaction even though the cost of mainframe was considered prohibitive.

With technical improvements and release of the latest Enterprise z server, mainframe is positioned to stay for a long time by enabling corporates to embrace latest industry trends, namely enterprise transformation (SAP on z/OS), cloud computing (z server is an ideal platform for hosting a cloud) and advanced data analytics (no doubts there). Far is not the day when we will have a pocket sized mainframe shipped to Mars to run the first corporate outside Earth.

Written by and posted on behalf of Rajeev Mohankumar

Posted by Prasanna Ramaswamy at 4:40 AM | Permalink | Comments (0) | TrackBacks (0)

Everyone has heard about the wide variety of tools available to support and encourage enterprise collaboration. But collaboration doesn't start with a tool, it starts with people. Collaboration is a hot topic right now, and has been for a good few years. The word 'enterprise' in front and the term becomes even more newsworthy. "Enterprise collaboration -- yes that's what we should be doing" is a sentence that can be heard in various guises across boardrooms all over the world. When you really think about it, collaboration is at the very heart of every business on the planet.  It's very rare that you find someone that is isolated from the rest of the company.  Most people are a part of a team that needs to work together to achieve the best possible results; that team is a part of many teams that all need to work together to help grow an enterprise.  We collaborate in pretty much everything we do at work, it's not always efficient and it's not always effective, but what if it were?

Frost & Sullivan along with Verizon Business and Microsoft conducted research around the impact of collaboration within the enterprise.  The results of the study showed that collaboration is a key driver of company performance (the study was conducted in 2006).  A global collaboration index model was developed which looked at variable factors that affect collaboration as well as several variables that affect company performance.

The highlight of the research project was that 36% of a company's performance was due to its collaboration index, 16& was due to strategic orientation and 7% was due to market and technological turbulence influence.  Here is how collaboration affected the various aspects of business performance:

Posted by SathishKumar Ramadoss at 3:33 PM | Permalink | Comments (0) | TrackBacks (0)

Cost-optimization is currently one of the high priorities for companies. It is very important to make sure focus does not shift from this during good times. According to a 2007 KPMG survey of more than 400 companies worldwide, 9 out of 10 cost reduction programs fail to achieve their targets, and the gains that are achieved appear to be short lived.

When a consolidation happens  in any industry, the market takes the news in awe. The market reacts swiftly initially with lot of news-spread, followed by strong analysis, view points and the market settles down accepting the new order and moves on. I’m sure the same will happen with the JDA i2 acquisition. JDA historically has been a great aggregator and have done acquisitions at strategic moments in its growth phase. Acquisition of Intactix, E3 and Manugistics all helped JDA solidify its position in the market, expand its footprint and establish itself as the leader.

Continue reading "JDA i2 acquisition: 3 point view" »

Posted by Kirthi S Subramanian at 2:15 PM | Permalink | Comments (2) | TrackBacks (0)

Continue reading "Enterprise 2.0 for Retailers" »

Posted by Shveta Arora at 8:23 PM | Permalink | Comments (0) | TrackBacks (0)