One Size Does Not Fit All for IoT
The mobile industry is betting big on the Internet of Things. Mobile network operators are looking to IoT to fuel their subscriber and revenue growth for many years to come. But since there aren’t many “people” out there who aren’t already connected, the next logical step is to connect “things.”
Things that could potentially be connected come in many different shapes and sizes; from small wearable devices up through modes of transportation – cars, trucks and even ships. Things can address different vertical markets, too – health care, sports, pets, automotive, industry, logistics and more. As a result, requirements differ greatly: Security, bandwidth, data prioritization, battery life, connectivity and intelligence all play roles in determining the proper solution.
To put the overall discussion in context, let’s touch on some of these requirements, starting with intelligence. Depending on the intelligence in the device, the requirements for bandwidth, security and battery life may be impacted. For example: Is the device able to go into sleep mode and only wake up at appointed times or when it knows there is relevant data to be sent, or does it stay in an always‐on mode and require constant pinging? In the latter situation battery life is greatly reduced. Constant use of bandwidth occurs as opposed to intermittent bursts with potentially more opportunities for malicious activities if the connection is always live. An intelligent IoT device is considered one that is capable of learning over time and being remotely controlled via software manipulations. This intelligence could play out when the device begins to learn when it needs to check and send data as opposed to being programmed for a set time. Additionally, remote sensors can be upgraded over‐the‐air when the appropriate type of software is enabled.
How much information is being sent and how often drives communications technology decisions. Even when considering the connected car, the vast majority of information exchanged today can be accomplished via 2G. But we know it’s not realistic to envision a 2G future. As 2G networks are phased out due to the high costs of maintenance and low service revenue for the operators, the solution must change. Additionally, the amount of data exchanged increases and the amount of sensitive information (such as health‐related sensor monitoring) drives the need for more sophisticated communications options. And then there are new technologies in the area of Wi‐Fi along with new alternatives such as LoRaWAN – long range WAN and SigFox.
The most complex “thing” in this discussion is the connected car. Think of it as an expensive smartphone on wheels, but with many more implications when it comes to securing the data and passengers. Let’s start with the personal data that currently goes through the communications network via a smartphone. It is pretty clear that it’s no different than your current smartphone scenario, but now let’s layer on car‐specific data. Today this information is retrieved and software functionality updated by a physical connection to the car. This trend is shifting to over‐the‐air functionality to enable faster identification and resolution of car reliability and safety issues. Although this traffic is encrypted and on a separate network in the car, that is going to change. In order to reduce costs, cars are moving to IP networks just like the rest of the communications and IT industry, introducing a greater opportunity for security issues. Although the case with all IP networks, the implications to a moving vehicle with passengers raises many concerns for what might happen. And when cars are communicating with each other and surrounding infrastructure – beyond the potential security issues – performance becomes extremely important as safety and road‐hazard data must be shared with almost no latency. Some experts set this goal at 1ms as the target for the acceptable latency for this information exchange.
Service providers must also consider not only the end‐user opportunity for new services through IoT, but also the implications to their networks that may not be in place today. In general, IoT places signaling and security challenges on the network not currently experienced. Some topics to consider:
- Signaling: Consider the possibility for a signaling storm caused by hundreds, thousands and even tens of thousands of devices trying to reconnect to a network over and over when there’s an issue. Signaling volumes on 4G LTE networks will grow at twice the rate of mobile data traffic through 2017, according to new research from Oracle. This is before IoT is added to the equation. Diameter signaling overload has brought down LTE networks in the past, so having solutions in place that can prevent it from happening will be critical.
- DNS Services: Today when a DDoS attack occurs, the strategy in place is to block traffic as a first line of defense. With IoT, that is not possible as blocking at the DNS level will also block the good traffic that still needs to get though. So there is a need for a highly scalable DNS solution that is capable of providing intelligent DNS protection.
- Firewalls: As LTE moves to sessions from connections and there are potentially 20‐30 sessions per end user on a single firewall, it doesn’t take long before the firewall could fail, as throughput is no longer the bottleneck for a firewall. The connection‐per‐second rate and the total number of connections it supports are going to be the key criteria for firewall selection in mobile core networks. In addition, the different M2M use cases will co‐exist with the traditional mobile broadband services for the consumer market. As a result, more sophisticated firewall policies will be needed in the mobile core to keep these services completely isolated and secure from one another.
- Encrypted traffic: Many IoT transmissions will be encrypted, especially when considering connected car‐ and health care‐oriented data. Content optimization is therefore not possible, but techniques such as TCP optimization, intelligent bandwidth control and QoS could help ensure the fastest possible data transfers for business critical operations.
Not to sound all doom and gloom for the future of IoT by any means, IoT opens a world of possibilities on all levels. Most importantly, the end users will have capabilities available to them they can’t even imagine today. Service providers have the opportunity to offer end users IoT and enterprise IoT services customized to specific verticals and functionality. Maybe cars will even have their own services in the future. And lastly, we cannot overlook the innovation that will come from the hardware and software vendors to enable the service providers to develop and deploy these new services.
F5 Networks provides carrier‐grade solutions that deliver multiple services on a unified platform to enable service providers to enhance quality of experience and generate new revenue streams. Learn more about F5 Networks tools to help service providers be ready for these challenges. www.f5.com