Browse by year:
Data Life Cycle
Ramki Pitchuiyer
Founder & CEO-Paloras
Tuesday, October 20, 2015
Life of Data
The last few decades have been an information revolution while the next few decades will be capitalization of the information revolution. Recently, there have been newly discovered ways to collect information, analyze it to answer questions, and report it in various forms. Now we can go even further in using this data to make predictions and also provide the intelligence to follow through actions on our behalf. The following paragraph highlights each step of this information era with examples in the industry, predictions on where we are headed, and how Paloras Corporation is helping in the Enterprise sector around customer experience initiatives.

Setting the Landscape
A little over a decade ago, we collected data because we had to. There were laws in place that dictated how much HR and customer data we had to keep. Around this time, we transitioned from documentation on paper to digital storage. Data was small but storage was huge: A friend of mine used to say an intern at Sun used one GB storage device as a coffee table! Now, we can store several gigabytes of data in a flash drive no larger than a paper clip.

Our technology at that time was focused on how effectively we could capture and store data based on RAS (Reliability, Accessibility and Serviceability) criteria. For archiving, we saved data in multiple physical locations to prevent loss. Retrieving data with speed became a priority, so we used metadata to describe the features of the information we collected. Also for speed, we had to ensure geographical proximity to data. Furthermore, we did not realize at the time that storing all data was necessary, and instead only kept data that we assumed was important and relevant. What we thought was important and relevant was based on what we believed were the questions that needed to be answered. Geographical proximity and commoditization of hardware led to rise in cloud services. With cloud, security decisions had to be made on what data is stored within an enterprise (or private cloud) and what can be sent outside of the enterprise (or public cloud). As more data grew in various dimensions of cloud, the technology matured enough such that accessibility and security became more important and reliability and availability is taken for granted. Many companies are going for hybrid approach with some applications (and therefore data) inside and some on cloud. The need keeps going up to understand and cross link the data within the hybrid framework and drive actionable outcome from the analysis.

Big data and Analytics
Interestingly, how we determine what data is relevant or not has changed. Today in this information revolution, all data is relevant regardless of its importance. We now understand that pre assumed questions we originally had are not the only ones that can be asked. We know the nature of breadth and depth of our questions varies over time requiring data that was originally considered not relevant to be relevant or even important now. We are now venturing into creating intelligent algorithms to analyze data. In real time, we want to, make conclusions, test hypothesis, formulate solutions, make connections with other data, take actions, and then do more of the same. We want to do this with least human interference, in lightning speed with highest amount of accuracy. Big data is piece of the puzzle. We still have challenges in combining various types of data structured and/or un-structured coming in from legacy systems and ever-expanding new systems. This is especially important when enterprises increasingly use hybrid (cloud and non cloud) application and data strategy. This is where a lot of products focus on.

Let's look at what this is with a simple example. Let's say we have a pre-assumed question "My budget is $5000 and I want to go on a vacation for 2 months. What is the best place given my constraints like Geography preferences, interest like culture or archeology, etc." To answer this, we are combining data from personal finance software, interests from TripAdvisor, expenses from Airbnb and generally comb the Internet through search engines to gather knowledge. Once knowledge is gathered, we put together various options. We then decide on an option, chart detailed plan. Executing the plan is even more extensive and may require additional set of tools.

Beyond Big data - Machine Learning
How do we move from providing answers to questions to predicting what questions we need to ask and the answers to them? For example, if the bigger question is "I want to go on a vacation. What are my options?" Wouldn't it be great that a system (or intelligence) helps pull in affordability from financial software budgets, time constraints across calendars, personal interests from social reporting and past habits and tell options one can choose? And based on intelligence gathered on these options, it can also highlight its pros and cons. Once a person picks an option, the intelligence works to do all the bookings and incorporates into schedules, financials and relevant systems.
Share on LinkedIn