The IoT Files – Infrastructure
IoT is a complex argument, we already know it. In my previous introductory post I tried to explain the privacy and security concerns that IoT is bringing to us (https://www.linkedin.com/pulse/iot-files-privacy-antonio-ieranò).
Most of those concerns are intercnnected one to the other, but have also a strong relationship with the next point: the infrastructure needed.
With Infrastructure I refer to a lot of things, that goes beyond the simply technical aspect, because a real IoT infrastructure goes way beyond the access protocol or the wireless.
Wireless Outdoorsindoors Connections
But since we named Access protocols and wireless stuffs, let us talk about the first easy infrastructure need: connectivity.
In a IoT world devies need to be connected, and just a few of them will be able to connect trough a cable connection.
We can imagine a cable connection to our SCADA environment, sure, but things would be a little harder if we consider our smartwatch or our autonomous driving car, I can’t think we can go with an Ethernet cable connected on those devices 🙂
Connectivity will be an issue in IoT for several reason, and sometimes I got the impression we underestimate the issue.
Let’s think about our homes, they are in most of the case connection unfriendly. It is not just the level of connectivity and bandwidth offered by our provider (that will be object of a later point).
We should start to design new home with wireless point in mind, probably, and enough network power to get all connected, but what about the old ones (that are the majority?).
At the moment the offering is still way far to be exhaustive, some steps with power-line and home wireless have been done, but just to mention the security and privacy concerns we mention before, this is still not enough.
The routers we use in our homes are all but efficient, and surely not able to deal with hundreds of devices not in terms of connection nor in terms of protection.
But is when we go out our homes things get harder.
What we should expect? a wireless coverage of urban areas is something we can imagine, but as we go out the urban area things get harder. Of sure we can use our phone provider that offer connectivity (at the moment through 344.5 G) but will this enough?
does every device use a sim and a contract? and what happen when we need to go out our city or our country? will roaming hit us down?
Infrastructure from this point of view is all but simple, we need to be able to transport petabyte of data (is what billions of device connected means, my friends) in a multinational context, providing access to different devices with different capability.
This will be a key factor that will make dramatically clear what is the digital divide. We are struggling with the digital divide so much today, the problem will be bigger in the IoT future, because if we are struggling to put in place the infrastructures now, how will happen when we will need bigger, faster, stronger and more complex infrastructures?
Large areas even in the more advanced countries will be cut out: countryside, mountain, islands…..
The promise to mitigate those concerns is called 5G technology. But I want to be clear, %G is far to be a complete standardized technology at the moment, it is still under development and most of the issues we named for IoT are issues also for 5G, from access to security to business models….
I personally think that 5G will be a useful step forward, but will not replace the heterogenic environment. wireless and ethernet will keep playing a big role and therefore interaction between the different technologies will be mandatory.
Like it or not even in the 5G world we will have to fight the bandwidth problem, because this will be the issue. data requires bandwidth and IoT means data, without data IoT does not exist.
Bandwidth is not a easy issue, because it means how to prioritize traffic, how to manage traffic coming form different sources..and so on.
the amount of bandwidth available and its management will be a key issue, it require clear infrastructures and models we still lack.
Even the government services will have to face the IoT revolution and become to be compliant, we cannot imagine that a life hyperconnected require a form manually compiled by the user, isn’t it? government infrastructure will have to shift dramatically towards a new model where informatization is not just a way to have more efficiency, but the only way to provide service.
It can not seems an infrastructural problem, unless you remember your experience when dealing with a government office…we lack of tools (HWSW), personnel, culture, policies, knowledge..isn’t this infrastructure?
While we will deal with new issues we should not forget we have a lot of old issues to deal with that can make hard the transition to the IoT.
Let’s name a couple that are so big (and so neglected) that I am wondering why we still talk about IoT.
Old Issues – DNS
Billions of device will try to connect to the internet, every device will look for partners to communicate. Unless we think all those device have hardcoded the partner address (wich is unlikely and highly impractical for a thousand of good reasons, one for all, flexibility) the device will need to translate a logical address to an IP address.
This service, nowadays, is done by DNS infrastructure. The DNS infrastructure is an area of big concerns, because it is subject to attacks, it is easily victim of geopolitical issues (a government closing the root, as an example) or poisoning entire zones for censorshipmass spying issues.
At the moment DNS around the world are really in a bad condition, mot of the carriers that offer DNS resolutions does not even size them proèperly, not talking about protecting. the reason is that DNS resolution service is not perceived as a key aspect, and it is not direct source of revenues.
If for a security perspective something is moving, with the DNS-SEC extensions, form a performance side this is still a pain in the butt. most of the time when you blame your provider for bandwith, if your page does not load is because of the poor DNS resolution service.
In a world of billions of devices this infrastructure, easy prediction, will collapse. Name resolution will need a support, what I am afraid is the developing of custom made legacy protocols (peer to peer style) that will address the problem in the lack of a commonly accepted solution, this will affect security and interoperability of IoT.
So if you think DNS is not a problem in IoT will be.
But DNS is a victim of a deeper problem, we all know that TCPIP v4 will not be able to scale to the IoT, but where we are not with TCPIP v6? let’s face the truth, we are still at the beginning, This is a big infrastructure concerns, because most of the infrastructure are not yet ready to move to IPv6, otherwise we will be already there.
There are big issues related to legacyold hardware, lack of knowledge form the technical people, absolutely not understanding form the decision makers that does not consider it an issue. so we are, in short terms, in a big sea of troubles at the moment.
When we talk about infrastructure we have always the same issue: who will pay for them?
we have to realize that all the needed infrastructure for IoT comes at a cost that someone has to pay.
Public and private will have to find a way to deal with this, because big investment will be needed.
Another painpoint is the timing: how long this infrastructure need to be set up? If prevision say we will have billions of devices for the 2020, I suppose those infrastructure will be ready for that time.
But wait, we are in 2016 now, and I can’t yet see thos ebig investment to cover and solve the issues we talked before. so may be the time will be an issue we will see sooner or later.
And we should remember that the infrastructures needed aren’t local ones, but international. Lack of standards, agreements will make it harder.
So we are seeing a big opportunity as well as a big headache.
but cheer up, as murphy told us, smile tomorrow will be worse.
next IoT files on Business models…..