Edge vs. Cloud in cities

We gathered up an A-team panel, led by Oren Gadot, Incubit’s CEO, which included Doron Gonen (NEC Israel Research Center), Ofer Shneyour (Applied Materials), Alex Yusopov (Axon Vision), Rami Anati (Pearls of Wisdom, now Elbit), and Avi Hadad (Tel-Aviv Yafo Municipality), representing a diversity of viewpoints regarding IoT implementations in urban, military and industry 4.0 contexts.

For those unaware of the debate held around this topic, a short explanation:

Devices collect data. This data can be sent in its raw form to a distant server, for analysis and further action. Alternatively, the analysis can be done on a processor on the device itself (“edge device”), and only the outcome of the analytical process is sent to the central location for further action.

For example: in an automated parking lot, operated by cameras capturing the license plate numbers, these two scenarios would mean either sending the actual footage to a command and control (C&C) center, where computers would run algorithms for licence plate recognition (LPR), and then match the license plate number with a phone number or a digital wallet identity, to complete the billing; or running the LPR algorithm in the processor of the camera itself, and sending just the license plate details to the C&C.

This may sound like a minor difference to most, but behind the scenes there are some important trends and considerations to bear in mind:

Devices are getting smarter and more powerful all the time. Our smartphones are also edge devices, and they are currently capable of shooting video in 4K with some impressing processing and storage power (and still, as opposed to the past, fewer people still store music or photos on the device itself, and in most cases they keep everything on the cloud).

Remember that one of the biggest constraints on such devices generating so much data, is that all this data needs to be sent somewhere over a network, and current bottleneck is bandwidth. Especially where there isn’t physical infrastructure such as fiber optics, and one needs to rely on wireless communications, mostly cellular – a pipe which is currently too narrow to fit all the data that needs to flow through it. Sending only the analytics rather than the raw data, means sending much smaller quantities of data over the network.

In some cases, the latency generated by sending the data elsewhere can be critical to certain processes, as we’ve learned from the case of Applied Materials, checking each chip they manufacture to detect defects, in a process compared to locating an ant in Manhattan within just one second (!!!).

Naturally, in many of the military applications, connectivity to a network just isn’t available in the field, and some of the processing must be done on site. Also, let’s not forget the sensitivity of the data, which could be compromised when sent to another location.

With all these arguments in favor of running processes on the edge, cities face a different situation, which is related to the cost effectiveness of any infrastructure project.

Compared to a factory production floor with a few dozens or hundreds of devices on average, cities deal with a different scale of hundreds of thousands of devices deployed, sometimes in areas which are difficult to maintain, on rooftops or high street light poles.

Therefore we must consider cost per unit (smarter devices with more technology on the edge, are usually more costly), maintenance (more electronics on the device compared to a “dumb” sensor usually means more malfunctions), and more power required to operate more complex processing on site, meaning a higher frequency of changing batteries for example, which is a burden on the maintenance team, also translating to higher costs.

Overall, it seems that we are looking towards a hybrid future, where cities will run many processes on the cloud, and heavy duty applications, mostly regarding high resolution video, would be either done on site, connect over fiber optics, or wait for higher bandwidth wireless connectivity (millimeter bands) to be deployed.[/vc_column_text][/vc_column][/vc_row][vc_row type=”in_container” full_screen_row_position=”middle” scene_position=”center” text_color=”dark” text_align=”left” overlay_strength=”0.3″ shape_divider_position=”bottom”][vc_column column_padding=”no-extra-padding” column_padding_position=”all” background_color_opacity=”1″ background_hover_color_opacity=”1″ column_link_target=”_self” column_shadow=”none” column_border_radius=”none” width=”1/1″ tablet_width_inherit=”default” tablet_text_alignment=”default” phone_text_alignment=”default” column_border_width=”none” column_border_style=”solid”][vc_gallery type=”nectarslider_style” images=”7470,7471,7472,7473,7474,7475,7476,7477″ bullet_navigation=”true” bullet_navigation_style=”see_through_ar_visualized” onclick=”link_image”][/vc_column][/vc_row]