In today's era of increasing connectivity, data is getting generated in vast proportions. Moreover, it is also important to be able to generate insights from it quickly and act accordingly. Gone are the days when one would move data into a data warehouse and then extract insights from it to act at a later date. Here are four scenarios why.
If a discussion around a brand is trending positively or negatively, that organization needs to take action then and cannot wait for a future time to do so. They might want to capitalize on the positive sentiment and amplify it or perhaps take action and remedy a trending negative sentiment. Both Twitter and Facebook provide several real time analytics capabilities leveraging big data technologies that they pioneered themselves. These analytics run within their cloud environment and provider users real time insights.
In order to analyze and generate insights quickly, it is imperative that we analyze the data where it is produced. Today, data is being generated via various sources including sensors, RFID tags, social interactions, feeds, web logs, user click streams etc. Let's now take the example of web logs. Organizations love to understand how users traverse through their website, which pages they visit next and so on to not only better design their websites but also to more effectively run marketing campaigns. A lot of these weblogs can be stored in Hadoop distributed file systems and analyzed. As several organizations today run their websites on cloud already and the weblogs are generated in the cloud, it is a no-brainer to do the analytics via Hadoop on the cloud itself.
Moving on to another increasing scenario where traffic lights are now collecting data on traffic patterns. Such data can either be analyzed while in stream via "stream processing" and then discarded, or stored in various data sources including Hadoop for further analysis. City governments can leverage this data to better predict traffic congestion in the future, re-route traffic at the time of congestion and so forth. Funneling the data to a cloud location might facilitate a common Cloud that could help other cities combine this with their data for even more accurate predictions. In addition, several cities might not have sufficient IT infrastructure in house for analysis of such large amounts of data. They could simply rent IT infrastructure from Cloud providers cost effectively.
Similar to the traffic scenario, Smart Water projects are also underway and reaping benefits. Sensors are helping capture data, and analytics is helping cities prevent wastage of this precious resource before it happens. Here is a map of all the Internet of Everything related projects from Cisco such as Smart Water, Smart Traffic and others that are in progress...
More and more mobile applications now typically have their backend processing in the cloud with developers focusing on the UI aspect of their application that sits on the mobile device. This approach is commonly referred to as 'Mobile Backend as a Service (MBaaS)' and is gaining popularity. These mobile backends need to run in the Cloud as they need to be elastic and be adaptive to changing demands as load on the mobile app can vary drastically over time. As a result, the data that is being collected from the mobile app is once again in the Cloud where the mobile backend resides, and consequently analyzing that data can be easily done there as well.
To wrap, let me leave you with another thought. As I have described above, several new sources of data to be analyzed are already in the Cloud. Organizations going forward might actually need to combine their data sitting in different Clouds to generate new insights. Cisco's Intercloud approach of connecting various Clouds will be an integral part of that next step.
For additional insights on this topic, do explore the following resources: