IoT Gateways: Designed out before designed in?
April 15, 2016
With the news that the lower end of the 4G LTE spectrum is to be opened up for low bandwidth M2M and IoT devices, with CAT1 and soon CAT0 connectivity...
With the news that the lower end of the 4G LTE spectrum is to be opened up for low bandwidth M2M and IoT devices, with CAT1 and soon CAT0 connectivity, one must consider if the much-anticipated IoT Gateway has obsolescence in sight before its widespread implementation. From the mobile networks, whilst this may first appear to be a huge investment they’re making in IoT, they actually have little choice.
With the well-publicized predictions of billions of IoT devices by the end of the decade, the 4G LTE infrastructure simply lacks the capacity for even a reasonable percentage of these devices to connect to the standard high-speed 4G LTE. Halving the delivered bandwidth can double the number of connections, and with typical IoT endpoint bandwidth requirements being mere kilobytes, a tiny fraction of that bandwidth is more than adequate – thus, by its inverse correlation, achieving the magnitude of connected devices that the industry will demand.
So why not force these ultra-low-bandwidth devices to use the relatively ancient and low-cost infrastructure of, say, 2G/GSM? In Europe this would work, at least for now, but in the United States, the biggest mobile carrier is still on schedule to shut down 2G by the end of this year, forcing migration to newer wireless technologies. Incidentally, other U.S. mobile carriers such as Sprint are looking to capitalize on this vacuum and continue to offer 2G to M2M customers, but realistically can they maintain that infrastructure long term? Who knows.
So now IoT endpoint devices can inexpensively connect directly to the cloud, bypassing the IoT Gateway. The question is do we actually want this? Arguably, there’s a parallel with Wi-Fi routers and 3G connectivity – yes, we can all get 3G within our homes, but I know very few who have used this to justify ditching their fiber broadband altogether. The reality is that even as the cost of wireless broadband has come down for consumers; for anything other than light users it is a more expensive option, with slower speeds and lower reliability.
IoT endpoint devices transmit/receive such little data they won’t suffer from insufficient accessible bandwidth, and adequate reception still isn’t ubiquitous, particularly in more remote areas. Surely security is the biggest risk here – a local infrastructure “tree,” where security can all be handled in the single trunk (router/gateway) and all branches (devices) function through that trunk permit a centralized location to manage what are now extensive security requirements. If we change that structure and every device has a direct connection to the Internet, the entire globe is rapidly becoming a single network that is only as secure as its weakest link.
For now maybe that’s not so problematic. The only companies developing such solutions are huge stalwarts of embedded security, and that primary need flows through the veins of every developer involved at every level. However, as this technology becomes mass market, once devices are being spun for literally cents for sale on our market stalls, cutting every possible corner in the process – then we have a real problem.