Big data - Hype on ! Hype off !
Hadoop hype has subsided recently although a major vendor like cloudera expecting to double their revenue this year (86% YOY) by selling Hadoop distribution. Though, at their current rete they are expecting operating profit in 2018. The major reason for this could very well be wearing off the hype which has been going on for a while.
As per the latest survey by Gartner, 54% of the firms don’t have any plan to invest in this area, in fact, if we look closely we would realize around 74% of the people are still thinking whether this would make any impact on their bottom-line. There are multiple concerns but none of them which cannot be taken care of. What’s interesting is that all this time new big data technologies has been growing and making an impact. For example, Cassandra and MongoDB have occupied the top position for NoSQL databases. These databases fit directly into the existing eco-system of the existing implementations although they are not born out of relational data model. What they offer is the scalability and performance at a cheaper rate.
Hadoop is Big data!
If you talk to a senior executive from any firm, they will invariably put big data in their strategy landscape. However, when we talk about the implementation of Hadoop things are different. Actually, here comes the major confusion, big data again is not all about Hadoop. One really has to understand that big data is a concept, it’s a relative term. 40 years back even 10 GB would have been a big data because at that time we did not have means to harness sense out of it in an economical manner. But in the current scenario my laptop can happily crunch this data. The crux of Hadoop or any big data tool is the scalability in an economical way. This concept is not new as we had MPP database like Vertica which actually is a columnar database but not really, lets park that discussion for some other time. Scaling up has been expensive but scaling out in 2000s is not at all expensive.
Big data and so on…
If Big data technology is like kings landing, Winterfell is IoT, they need each other!! If we really look at the various trends coming in, we would realize that cloud will have a lot to offer. Many data intermediaries would be coming into the picture providing data services and taking cloud to a totally different level. The current market of big data is assisting technology advancement and bringing open source to a totally new level. Personally, I feel Hadoop has a long way to go because still we don’t know what sort of data deluge we will be experiencing in coming time, with respect to current scenario NoSQL fits better.
Yes, even i believed that but the firms are being very cautious and not really very sure how then can extract the best out of it. In fact the recent surveys (like gartner) has indicated this aspect. This is exactly what i am talking about, NoSQL is currently fitting better in their current landscape...but lets see, its a dynamic market
MS in some of its presentations confirms that big data will grow for no other reason that its free licensing. All valid points but I still feel that almost all big enterprises would move to Big Data Systems like Hadoop over the next few years, if not for reporting but for Analytics... If not hosted in their own data centres then may be through 3rd party service provider in different level of services from IAAS, PAAS or SAAS...