Bottom Line: Bad actors quickly capitalize on the wide gaps in machine identity security, creating one of the most breachable threat surfaces today.
Why Machines Are the Most Challenging Threat Surface To Protect
Forrester’s recent webinar on the topic, How To Secure And Govern Non-Human Identities, estimates that machine identities (including bots, robots and IoT) are growing twice as fast as human identities on organizational networks. Forrester defines machine, or non-human, identities as robotic process automation (bots), robots (industrial, enterprise, medical, military) and IoT devices.
The webinar points out that one of the fastest-growing automation types is software bots, with 36% used in finance and accounting, 15% used in business line and 15% in IT. The webinar also points out that in 2019, there were 2.25 million robots in the global workforce, twice as many as in 2010 and 32% of global infrastructure decision-makers expect their firms to use robotic process automation (RPA) over the next 12 months.
According to the Forrester Consulting white paper, Securing The Enterprise With Machine Identity Protection, over 50% of organizations find it challenging to protect their machine identities today. Unprotected machine identities are making it easy for bad actors to take control of entire networks of devices. Bad actors rely on organizations’ bots to provide the cover they need to attack networks and devices, often undetected for months or years.
Forrester found that machine identities are left exposed to bad actors because organizations aren’t adopting the tools they need to create and manage a centralized Identity Access Management (IAM) strategy across all machines. This includes defining and enforcing policies, auditing each machine and endpoint and better integrating support across machines and monitoring systems.
Furthermore, by adopting a more modern Privileged Identity Management (PIM) approach, organizations could solve many of these challenges. Leading PIM solutions providers include Centrify, which has succeeded in adapting to the ephemeral nature of securing machine identities by delivering machine identity and credential authentication based on a centralized trust model.
The Forrester report’s bottom line is that machines are isolated, exposed and more vulnerable than any other endpoint on a network. The following graphic compares protection strategies and finds a majority of organizations struggling to deliver them:
Machine Identities Are Networks’ Weakest Security Link
According to a Venafi study, machine identity attacks grew 400% between 2018 and 2019, increasing by over 700% between 2014 and 2019. Malware capable of compromising machine identities continues to gain momentum, doubling between 2018 and 2019 and growing 300% over the five years leading up to 2019. According to Kount’s 2020 Bot Landscape and Impact Report, 81% of enterprises are regularly dealing with malicious bots today and one in four say a single bot attack has cost them $500,000 or more. Furthermore, many organizations may not realize how many bots and machine identities they have – and bad actors capable of creating hundreds using automated scripting tools.
Forrester provided the following data points underscoring how vulnerable machines are to botnet and identity-based attacks today:
- The 2017 Mirai botnet attack is a cautionary tale of the dangers of using default security credentials on machines and IoT devices. Using botnets to automate scans of vast blocks of IP addresses for potential telnet ports to log into, the Mirai botnets were programmed to rapidly try a series of basic usernames and passwords to gain access to IoT devices and machines. The Mirai botnets were successful, gaining control of thousands of machines and orchestrating them to deliver one of the largest DDOS attacks in history.
- It’s common for enterprises to lose track of how many bots they’ve created, giving malicious actors the perfect cover to mask their movements. Instead of creating their bots, malicious actors look to disguise their movements across a network with a company’s bots. Forrester’s webinar mentioned how a large North American insurance provider deployed 400 software bots for customer-facing digital chatbots and processing claims, among other tasks.
- There’s often no oversight of who has the rights to create and launch bots internally, leading to potentially thousands of bots without secured identities. One of the most troubling findings presented during the webinar is how loose the process is to create a bot – with no checks and balances in place or means of achieving consistent identity management.
How To Strengthen Machine Security
The more challenging any machine threat surface is to protect, the more opportunity it provides bad actors to breach them. A good place to start is by clarifying who owns keeping Transport Layer Security (TLS) and previous-generation Secured-Sockets Layer (SSL) client and server certificates, code signing certificates, Secure Shell (SSH) host and cryptographic keys so they are kept up to date. Letting those fall through the cracks will leave thousands of machines exposed and exploitable on networks.
Prioritizing machine identities and securing machine credentials is a must-have in 2021, as botnet attacks are quickly increasing due to bad actors’ being able to spin up thousands of them in days. The following are key steps to get started:
- Taking a Zero Trust approach to managing every machine identity authentication on a network now could save thousands of hours and dollars in the future. Taking a least privilege access approach to managing machines now will pay off in the future, as the workloads of machines and non-human entities continue to grow more complex. The Forrester webinar expands on this point by explaining how new, more complex inter-machine relationships are evolving quicker than legacy approaches to endpoint governance and security can keep up.
- Privileged access controls need to be more adaptive, secure and scalable than many organizations’ static-based approaches to securing machines are today. Forrester recommends replacing long-standing hardcoded credentials with session-based ones assigned via API calls from a vault. Machines are being used 24/7 and have access patterns completely different from humans using the network, making dynamically-assigned, ephemeral credentials even more important to protect a network. Privileged Identity Management (PIM) proves effective at providing privileged access controls for machine identities, with Forrester mentioning Centrify, HashiCorp and others as leaders in this area. Centrify’s approach is noteworthy in enrolling machines with its platform via a client to establish a trust relationship, so applications running on that machine can also be authenticated using a short-lived, scoped token.
- Monitoring more machines on a network often leads to a transition from legacy to integrated log monitoring systems that can capture, analyze and report anomalous activity across a network. Log Monitoring systems are proving invaluable in identifying machine endpoint configuration and performance anomalies in real-time. AIOps is proving effective in identifying anomalies and performance event correlations in real-time, contributing to greater business continuity. One of the leaders in this area is LogicMonitor, whose AIOps-enabled infrastructure monitoring and observability platform have proven successful in troubleshooting infrastructure problems and ensuring business continuity.
- Perform periodic audits to track all bots and machines in use across an organization, using Microsoft Active Directory to inventory and manage all of them. One of the most valuable take-aways from the Forrester webinar is the need to manage machine identities and their credentials centrally. Forrester mentions Microsoft Active Directory as one option. The companies providing services in this area include Centrify, which pioneered Active Directory bridging to authenticate human and machine identities based on a centralized model from a single identity repository.
Machines, or as Forrester calls them in their webinar, non-human identities require more precise, adaptive and ephemeral identity structures and access controls. CISOs and CIOs need to take greater ownership of machine identity authentication and provide Identity Access Management (IAM) and Privileged Access Management (PAM) down to the bot and non-human identity level. With the exponential growth of malicious bots tracking machine identities, now is the time to place machine identities among the highest priority of any cybersecurity strategy in 2021.