With Web applications gaining prominence in the enterprise technology space, Biztech2.0 caught up with Bhavin Turakhia, CEO, Directi, who shares his insights here about a scalable architecture for Web applications.
How can an enterprise build its own scalable architecture? What should a CIO’s priority be while embarking on to this?
To build a scalable architecture, an enterprise has to first determine what applications are being developed. Targets such as scalability, optimal utilisation, responsiveness, availability, cost and maintenance effort need to be kept in mind while trying to build an architecture for Web applications. In order to keep a balance, the first four targets should be maximised and the last two should be minimised.
The above mentioned are targets that a CIO seeks. Variables that are needed to tweak those targets are platform selection, hardware changes, application design, deployment architecture and monitoring mechanisms. All these variables are involved in defining and creating a scalable architecture. As a CIO looking at applications, one has to bear in mind how many users are going to be served by how many applications, type of availability, type of responsiveness and uptime. It is essential that these variables are communicated to all those who are responsible for developing a particular platform.
What are the various platforms that could be used to create this architecture?
The choice of platform will depend on the decision taken by the CIO with regard to the development time and the speed of the application. A compromise has to be made between these two. Most platforms are scalable but then again one has to weigh the pros and cons to get the perfect one. Important decisions while building a scalable architecture would revolve around the language that would be used for coding and the database and cache that is would be used.
Legacy information and database systems are being migrated to the Web, what is your take on this?
Many companies still have legacy applications but a lot of them have been migrated to Web-based technologies, J2E, SOA and enterprise technologies. I don’t really see any challenges but some of the elements that one would need to keep in mind are that legacy applications are typically two-tiered applications while the Web has a three-tiered type of application architecture. Going further, SOA entirely talks about loosely covered architecture, which has various end-points exposing different Web services that are consumed by numerous users. Theoretically, one could look at mashups and the like.
When you look at the interface, data could be coming from multiple sources through architecture such as SOA and then it is all collated on the user interface (UI). Earlier legacy applications used to have text-based interfaces, now there is a lot of processing and validation that is possible on the UI side itself, which means that when you are migrating older applications to the newer paradigm, you would want to leverage abilities such as RIA (rich Internet applications) that can perform numerous activities. All this provides a holistic experience to the user.
Earlier, a lot of intelligence used to be on the server side but now the intelligence is being divided from both the client side and the server side. That is an area that needs to be kept in mind while migrating legacy systems. Some of the other challenges that exist include downtime as a lot of these applications are mission critical and migration typically requires some kind of downtime. When you are moving to a new data structure, it must be ensured that the data integrity is not lost. Application developers and vendors will have to try and come up with solutions that minimise the downtime during migration.
What trends do you see in the Web applications space?
If you look at Web2.0 and now Web3.0, which is called the semantic web, there is no clear demarcation between the two. Web2.0 applications talk about usability and interface. Hence, a lot of emphasis is laid on UIs and user experience. Earlier web pages used to be static and refreshed but now they are far more interactive. The whole Web page doesn’t need to be refreshed; only the area that needs to be accessed is refreshed.
Web 3.0 is the semantic web. HTML was created as a mark-up language for display but it doesn’t give any semantic data inside a particular page. Web pages have so much information that is currently looked at in a non-semantic way. I see these trends as gaining prominence.
Please talk about Directi and the solutions that it provides.
Directi develops mass-market Web products. The combined product portfolio of various Directi businesses includes - communication and collaboration apps, social networking software, Instant Messaging, context analysis engines, antispam and antivirus solutions, large scale billing and provisioning platforms, traffic monetisation solutions, online advertising solutions, DNS management products, Linux and Windows Server Management software, Web hosting control panels and much more.
Talking about security, what major threats do you see infesting the Internet that could directly harm enterprises?
As of now, the major threat on the web today is distributed denial of service or DDoS attacks, which are very difficult to identify and prevent because they are typically conducted by malicious hackers.