IoT Security Use-case – Part 3 – Dependency Computing

Share

Welcome to our four-part IoT Security Use-Case series! Here’s how we’ve broken things down so far:

Part 1, The Challenge, highlights a use-case for a financial organization whose business strategy was based on replacing expensive bank branches with Automatic Teller Machines (ATM) and Interactive Teller Machines (ITM). They chose this as their growth strategy because branches are limited to specific locations, slow to roll out and expensive to outfit. On average, branches take months at time, sometimes close to a year to turn up. The IoT based ITMs can do 95% of what a customer needs including allowing them to interact with a human. All the while, the IoTs can be deployed in a matter of days or weeks.

Part 2, IoT Security Fundamentals, lays out the necessary functions required for securing purpose-built technologies. Especially when they need to operate on a number of distributed public or private networks. And purpose-built technologies don’t have the required resources to self-secure.

We’re now at Part 3, where we will outline why traditional security approaches just can’t secure an IoT platform of this type. Part 4 will cover the solution to this complex security scenario.

For Part 3, let’s start by breaking down the components of an IoT application Ecosystem before we get into IoT security. Securing IoTs is not limited to securing only the IoTs themselves. IoT platforms function in ecosystems that are made of not just IoTs, but one or more remote applications that are operated by one or more vendors.

For our ATM network scenario, the ecosystem includes a banking ledger application running in a colocation data center. A monitoring application running on Amazon AWS using a different set of instances in a dedicated VPC. A SaaS application providing 24×7 physical security surveillance service the bank has contracted. As well as an Authorization-as-a-Service provider the bank uses to process external transactions.

Then there are the Teller Machines. There are several types of Teller platforms that include traditional ATMs and two different Interactive Teller Machines (ITM) types. A unit with a smaller footprint and a larger unit with greater cash holding capacity. The ATM / ITM IoTs are distributed across many cities, placed in a variety of locations and location types from office buildings, stores, malls, courtyards and airports. The systems connect via a variety of Internet connection types that include LTE service, Internet WiFi service and Ethernet connections from the local facility.

In this scenario, using traditional security tools, each platform requires completely different types of security tools to perform the various security functions for the various platform types – cloud, SaaS and the different IoTs. This means that each cloud instance, each SaaS, and each IoT require a different type, batch and brand of security tools. And each different security infrastructure needs to provide access, application and content control, threat management, privacy and identity for each of applications banking and monitoring applications, another for each ATM and yet another for each ITM.

This could add up to over 24 different security tools – If the tools that provide the different functions defined above actually existed. In many cases, especially with IoTs, all of the necessary tools simply don’t exist or don’t exist consistently for the different platforms. Here is a breakdown of the security options actually available:

  • The ATMs did not have any onboard or commercial security options.
  • ITMs do have support for Access Control and Privacy but nothing else.
  • The Cloud Applications do support the full spectrum of security, but require multiple disparate technologies that have a very convoluted implementation and data flow.
  • The SaaS applications have no meaningful security options. In many instances organizations opt to use VPNs or use an encrypted connection but ultimately have to trust the SaaS provider for all other security functions.

This approach is considered perfectly reasonable today – and it is absolutely insane. The number of different technologies coupled with the complexity of acquiring, implementing, operating and refreshing each different tool is an expensive and resource intensive way of getting marginal security. All-the-while assuming and managing risk for some parts of the platform because the security functions required just don’t exist for all platforms. Furthermore, the ones that do exist are inconsistent in how they apply security.

This creates complexity. Significant complexity. And — Complexity is the enemy of security. The complexity of managing the many policies, technologies, products, vendor relationships and integrations between the various technologies creates insecurity and drives organizations to spend more time managing products than security. Hence spending more and more on security does not always render the desired results.

It’s fair to say that the more security tools that are implemented, the more complex the security will be. And it is complexity that creates gaps and makes you less secure.

However, there is another factor to consider as it relates to traditional security – even if an IoT or application is operating on a shared network that is protected by the latest and best security tools; and even if they are designed specifically for IoTs; the IoT platform will neuter them. 

This is not a matter of better or different tools, the traditional security model is broken!

Dependency Computing: A New World of IoT Security Challenges

The security model is broken because how we compute has changed dramatically. IoTs use a compute model called dependency computing. With dependency computing, the IoT is dependent on the application and the application is dependent on the IoT.

Consider the impact of dependency computing on IoT security.

In a shared network, where multiple IoT brands exist (think smart thermostat, TVs, Fridge, smoke detector, etc.), each IoT brand has a dependency, and by virtue of that dependency, a connection to a different application that is 1) remote, 2) operates in the cloud, 3) is controlled and managed by a third-party, and 4) has privileged access to one or more IoTs that operate on your network.

IoT brand A is dependent on and connected to application A. IoT brand B is dependent on and connected to application B. IoT brand C is dependent on and connected to application C, and so on. It would not be unreasonable to foresee an organization using hundreds, if not thousands, of IoT brands in the next few years.

Each application that an IoT on one network is connected to is also connected to countless other IoTs on different networks. And that application has privileged access to all of these IoTs – often over the public Internet. To complicate matters further, many IoTs are also remotely managed, either directly or via their application via another set of devices.

This creates a platform triangle in which many distributed IoTs and the application or applications managed by remote devices are interconnected and dependent on each other. The risk and exposure of this model are numerous in the event of a compromised IoT, application or device. These include:

  • In a case of compromised applications, especially those accessible over the Internet.
  • The compromised application has privileged access to the IoTs and can use the existing privileged access to scan and capture communications on the network the IoT operates on.
  • The compromised application can also be used to fully compromise the IoTs on one or more customer networks, allowing the attacker further access and control.
  • With compromised IoTs on a network, the attackers can:
    • Denial-of-service other devices, systems, application or platforms;
    • Inject manipulated data that can be fed to various systems;
    • Compromise other systems on the network otherwise known as cross contamination.
  • Attackers can also gain access through compromised IoT management, especially mobile phones.

Denial-of-service, data manipulation or compromise of other systems – including other IoTs – could impact critical systems such as:

  • Medical devices such as infusors, ventilators, respirators, or monitors for various body functions.
  • Vehicle and transportation system functions such as the car mechanical sensors, engine and transmission functions, braking system, navigation, infotainment systems However, drones, aircraft and ships with many critical functions are also vulnerable.
  • Building systems such as HVAC, elevator controls, and life safety systems.
  • Financial systems like the credit card machines, ATMs and ITMs described here.
  • Critical infrastructure control systems such as electrical grid, dam controls, air traffic controls.
  • Supply chain and manufacturing platforms that incorporate aspects of the platform from raw resources to retail.

Social Apps Plugins and Integrations

Another dynamic in this equation is social media integration, including apps and plugins. Systems like Facebook, Spotify, Pandora, Google, LinkedIn, and Amazon should be recognized as spyware, albeit sanctioned spyware. IoT Social media integration, though it may apply to only a portion of IoTs used in organizations, can significantly convolute how they are secured.

As you may have already concluded, the IoT application platform ecosystem is a tangled web of interdependencies between distributed devices on many different networks, using remote applications that are operated and controlled by a third-party, both of which are managed by one or more users with a variety of further devices that can connect from anywhere.

The IoT dependency computing model has many, many parts owned and operated by multiple parties, each with no visibility or control over platform elements they don’t own. The network the IoT runs on may be owned and operated by one party, the IoT itself by another, the application(s) by yet another, all of which may be managed by a different third-party.

Any security implemented between the many interconnected parts of the IoT ecosystem ultimately has no teeth. The security of one party impacts the security of all other parties!

With today’s remote cloud and SaaS applications, along with the pandemic rate IoTs are infiltrating every use-case, the concept of fence and gate security for a bunch of devices on the same network is naive, soon to be negligent. It’s no longer about better tools, it’s about a better model!

Stay tuned for Part 4 – Securing IoT, the Solution…

 

About Acreto IoT Security
Acreto IoT Security delivers advanced security for Crypto-IoT Ecosystems, from the cloud. IoTs are slated to grow to 50 Billion by 2021 and are on track to be the biggest consumers of Blockchain and Crypto technologies. Acreto’s Ecosystem security protects Crypto / Blockchain and Clouds as well as all purpose-built IoTs that are unable to defend themselves in-the-wild. The Acreto platform offers simplicity and agility, and is guaranteed to protect IoTs for their entire 8-20 year lifespan. The company is founded and led by an experienced management team, with multiple successful cloud security innovations. Learn more by visiting Acreto on the web at https://acreto.io or on Twitter @acretoio.

Babak Pasdar
Babak Pasdar
Babak Pasdar is an ethical hacker and a globally-recognized expert in Cyber-Security, Cloud, and Crypto-currency. He has a reputation for developing innovative approaches and methodologies for the industry’s most complex security problems. Before Acreto, Pasdar brought the first proxy-in-the-cloud platform to market, even before the word “cloud” was coined. He called in security in the "Grid". Named one of New York’s Top Ten Startup Founders over 40, he has built and successfully exited two Cyber-Security technology companies and his innovations have been widely adopted by the industry.

Watch Video

Replay





Interested In ...

Show Buttons
Hide Buttons