|
Peer-to-Peer
How Rediff.com got its groove back
Plagued by data overflow and having no way to keep its active
and inactive data segregated, Rediff went in for ILM. By Tanu Talwar
With
online portals growing at a stupendous pace, the sustainability of any portal
is determined by the user experience that it guarantees for which enhanced application
performance becomes a necessity. In order to ensure that a user gets the best
online experience, it is imperative that a site offers information and services
at the click of a button. However, to bring about efficiency in its performance
its imperative that the site is also equipped with an efficient IT infrastructure.
Rediff.com, is a leading provider of Internet news, information, communication,
entertainment and e-commerce services. As its user base grew, the company faced
the problem of managing enormous data volumes and the user experience at Rediff
began to deteriorate. Venki Nishtala, CTO, Rediff, says, Being a consumer
portal we have to deal with large volumes of data that grow every minute. These
records are of immense importance and need to be continuously stored and updated
for future reference.
Besides managing and storing entertainment and other information, the portals
e-commerce services required the site to handle confidential customer and seller
information that had to be vetted and made available to clients every now and
then.
Performance degradation
Rediff had implemented several data intensive custom applications with multiple
instances, for various online functions. These applications were designed in-house
and they did not offer features such as archival, purging or sub-setting resulting
in huge storage and data management problems. Being a B2C site, the portal had
to deal with numerous transactions captured from multiple applications and distributed
across instances. Lacking a structured storage process, locating these transactions
was a major hurdle. Then again, with no differentiation between operational
and production data, separating active from inactive data was tedious, time
consuming and a direct drain on performance. Nishtala explains, Over the
years, huge volumes of data stored with us had become inactive due to infrequent
usage. However, as there was no systematic separation between operational and
unused data a lot of valuable time and storage space was spent each day in taking
backups of both used and unused data.
Moreover, with the database burgeoning the performance of in-house designed
custom application was adversely affected. The performance of functions within
this application began to deteriorate causing it to run slow. This application
degradation in turn increased the backup time as well as the storage costs.
Nishtala asserts, Even though we had large volumes of records and data
that were no longer in use they could not be deleted. These records contained
valuable information that might be needed in the future. In order to avoid
any contingencies arising in the future, the portal was not only spending resources
in storing unneeded information but it was also incurring millions of rupees
of expenditure to pay off mounting IT maintenance bills.
However as the performance degradation started to hamper user experience resulting
in the bottom line suffering, Rediff decided to take quick action. Nishtala
states, Our commitment to an excellent user experience requires that all
data captured be available for online access 24x7 and for 365 days. However
the tremendous increase in the volume of data had become a major concern for
us as it was acting as a hindrance in meeting our standards and performance
expectations. In order to transform its overall operations, the company
decided to go in for a well-structured Information lifecycle management (ILM)
strategy besides, reducing the load on its production database by relocating
e-commerce transactions. The performance issue forced us to evaluate the
option of moving the historical data from the production environment,
adds Nishtala.
Company: Rediff
Industry: Media and entertainment
Solution: Solix ARCHIVEjinni
Implementation partner: Solix Technologies
Time taken for the implementation: Eight weeks |
Choosing Solix
Having decided to do something about its IT woes, Rediff
went searching for a solution to upgrade its deteriorating applications. The
company considered a number of ILM vendors before deciding to go with Solix
ARCHIVEjinni. Sai Gundavelli, CEO, SOLIX Technologies explains, Rediff
required an organized ILM function that would help it manage and serve information
by assigning a type or category based on access rules, retention requirements
and business practices. The key was finding a solution that could help
Rediff automate the process of selection, archive, purge and focused extraction
of data. Moreover, it was essential to ensure data integrity as well as online
access to archived data and non-production instances to make the project a success.
From Rediffs perspective there were a number of factors
in favour of Solixs solution. Its patent pending interoperability layer
technology, enables companies to maintain transparent read-only
access to archived data while ensuring data integrity remains relationally intact,
even after upgrades. The result is a secure production environment and streamlined
performance. The solution also offered to address the storage cost by eliminating
unnecessary data duplication with powerful instance subsetting capabilities.
It ensures enhanced application performance by relocating inactive data from
the production database to a secure online data archive.
| Rediff.com is an online provider of news, information,
communication, messenger, mail, entertainment, e-commerce, shopping, auctions,
sports and other business services. The site provides a platform for millions
of people across the globe to connect with each another. Founded in 1996,
Rediff is headquartered in Mumbai (India) with offices in New Delhi and
New York (USA). It has over 45 million users growing at 20 percent quarter-on-quarter
and over 165 advertisers. |
The implementation cycle
|
To transform its overall operations,
the company decided to go in for a well-structured Information lifecycle
management (ILM) strategy besides, reducing the load on its production
database by relocating e-commerce transactions
|
The solution was implemented by Solix over eight weeks. A two-member team from
Solixs professional services group closely studied the portals requirements
and demonstrated the most effective way to tackle the situation. The team along
with the Rediffs engineering team planned out a roadmap to discover the
metadata of the custom application that had been built in-house and based on
that built the configurations for data archiving. Challenges faced by the team
included finding a way to relocate the bulk of database transactions without
compromising on online data access. It was determined that such a complex task,
taking into account the intricacies of data models stored in multiple custom
applications deployed at Rediff, would take considerable time and there was
always the danger of loss of data integrity and access.
Taking these aspects into consideration, Solix presented a strategy for archive
and purge as well as instance subsetting for entire online applications including
the order management system (OMS). The application consultants and the users
at Rediff finalised the details of data retention, extraction and data storage.
The project was divided into two phases. The first phase involved archive &
purge and the second phase was about instance subsetting. Solix further implemented
its configurable metadata driven ILM tool ARCHIVEjinni at Rediff using its in-built
features such as the configurator and MiniMe. The team from Solix also trained
the personnel at Rediff in the use of Solix ARCHIVEjinni.
Enhanced application performance
Post-implementation Rediff got its return on investment. The solution provided
Rediff the capability to design configurations for seven custom applications,
which enabled automatic archival and purging of transactional data into separate
instances. With the solution in place, Rediff is now capable of providing data
subsets for testing purposes, instead of backing up all its data on a daily
basis. We have achieved around 75 percent reduction in storage requirements.
Now we keep a few months data on the production server and the rest in online
archiving storage. We are experiencing more than a 50 percent Performance mprovement
in our online services after implementing ARCHIVEjinni, adds Nishtala.
The deployment has empowered Rediff to accommodate any changes in its applications
and data retention needs even with minimal support from the solution provider.
Besides saving on huge storage and maintenance expenses, the organisation has
been able to minimise its storage, backup and recovery time.
|