Data Analysis: Edge vs Cloud

The IoT intelligence can be seen at three levels: devices, gateways and centralized computing.

The first two are in the Edge and the third is in the Cloud, we will see that analysis will be convenient to the Edge (much less developed so far) and which in the Cloud.

Some factors that can help the development of analysis on the Edge are:

• Speed, capacity , allow to create a high  speed input for analysis and action, cloud applications rely on network speed and the capacity analysis processing large amounts of data, putting more analysis on the Edge these two factors would be reduced.
• Security with data encryption, you can help meet safety requirements and regulatory issues in each country, an issue that has not been discussed much, but should be settled in the IoT and has an easy solution from the Cloud.
• Cost . - Although the devices and gateways Edge will be more expensive as having more or less intelligence, they can save money and be more cost efficient because they require less data transport and storage, as well as reducing maintenance costs and the expense of critical battery basic in sites with difficult access.

Factors favoring Cloud:

• Efficiency, scalability . Movin the analysis to the cloud is more efficient, developments are reused for multiple sites, the load SW is unique and scalability and growth is favored.
• Computing power , a centralized system comes with increased computing capacity and its evolution and expansion is much simpler and faster.
• Learning. - You learn a variety of networks, multiple sites and testing can be done on different heterogeneous networks.

Trends in the future

• There will be more data transmission than the current capacity, we are seeing than some providers of analytics solutions are developing their products in the cloud, if estimates data traffic are met, the transmission capacity of data networks will become a bottleneck, teh need of extension of the network could put undue cost to derail the IoT, as well as IoT solutions go success the data transmission will multiply faster than the potential growth of networks.
• Data stored in the cloud can reach dimensions that analysis applications can not treat and analyze them, so decentralization in the Edge where you can decide what data to analyze, transfer to the Cloud or disposal must be the architecture future.
• In the early stages of IoT where we are, the focus is on connectivity devices rather than their intelligence. But in future systems especially the Edge will analyze the value of the data, and generate data which is transmitted.
• In the IoT there will be multiple and heterogeneous sensors, machines, devices and applications on the Edge, to harmonize and have a robust process should have a record of transactions and a system validationfor them, that provides security and confidence.