With the prevalence of always-on connectivity and advancements in technology that is available today, threats are evolving rapidly to exploit different aspects of these technologies. Any device is vulnerable to attack, and with Internet of Things (IoT) this became a reality. In October 2016, a series of distributed denial-of-service (DDoS) attacks were launched against DNS servers, which caused some major web services to stop working, such as GitHub, PayPal, Spotify, Twitter, and others [1]. Attacks leveraging IoT devices are growing exponentially, according to SonicWall, 32.7 million IoT attacks having been detected during the year of 2018. One of these attacks was the VPNFilter malware.

This malware was leveraged during an IoT related attack to infect routers and capture and exfiltrate data.

This was possible due to the amount of insecure IoT devices around the world. While the use of IoT to launch a massive cyber attack is something new, the vulnerabilities in those devices are not. As a matter of fact, they've been there for quite a while. In 2014, ESET reported 73,000 unprotected security cameras with default passwords [2]. In April 2017, IOActive found 7,000 vulnerable Linksys routers in use, although they said that it could be up to 100,000 additional routers exposed to this vulnerability [3].

The Chief Executive Officer (CEO) may even ask: what do the vulnerabilities in a home device have to do with our company? That's when the Chief Information Security Officer (CISO) should be ready to give an answer. Because the CISO should have a better understanding of the threat landscape and how home user devices may impact the overall security that this company needs to enforce. The answer comes in two simple scenarios, remote access and bring your own device (BYOD).

While remote access is not something new, the number of remote workers is growing exponentially. Forty-three percent of employed Americans report spending at least some time working remotely, according to Gallup [4], which means they are using their own infrastructure to access a company's resources. Compounding this issue, we have a growth in the number of companies allowing BYOD in the workplace. Keep in mind that there are ways to implement BYOD securely, but most of the failures in the BYOD scenario usually happen because of poor planning and network architecture, which lead to an insecure implementation [5].

What is the commonality among all the technologies that were previously mentioned? To operate them you need a user, and the user is still the greatest target for attack. Humans are the weakest link in the security chain. For this reason, old threats such as phishing emails are still on the rise. This is because they deal with the psychological aspects of the user by enticing the user to click on something, such as a file attachment or malicious link. Once the user performs one of these actions, their device usually either becomes compromised by malicious software (malware) or is remotely accessed by a hacker. In April 2019 the IT services company Wipro Ltd was initially compromised by a phishing campaign, which was used as an initial footprint for a major attack that led to a data breach of many customers. This just shows how effective a phishing campaign can still be, even with all security controls in place.

The phishing campaign is usually used as the entry point for the attacker, and from there other threats will be leveraged to exploit vulnerabilities in the system.

One example of a growing threat that uses phishing emails as the entry point for the attack is ransomware. Only during the first three months of 2016, the FBI reported that $209 million in ransomware payments were made [6]. According to Trend Micro, ransomware growth will plateau in 2017; however, the attack methods and targets will diversify [7].

The following diagram highlights the correlation between these attacks and the end user:

Figure 1: Correlation between attacks and the end user

This diagram shows four entry points for the end user. All of these entry points must have their risks identified and treated with proper controls. The scenarios are listed here:

• Connectivity between on-premises and cloud (entry point 1)
• Connectivity between BYOD devices and cloud (entry point 2)
• Connectivity between corporate-owned devices and on-premises (entry point 3)
• Connectivity between personal devices and cloud (entry point 4)

Notice that these are different scenarios, but all correlated by one single entity: the end user. The common element in all scenarios is usually the preferred target for cybercriminals, which appears in the preceding diagram accessing cloud resources.

In all scenarios, there is also another important element that appears constantly, which is cloud computing resources. The reality is that nowadays you can't ignore the fact that many companies are adopting cloud computing. The vast majority will start in a hybrid scenario, where infrastructure as a service (IaaS) is their main cloud service. Some other companies might opt to use software as a service (SaaS) for some solutions. For example, mobile device management (MDM), as shown in entry point 2. You may argue that highly secure organizations, such as the military, may have zero cloud connectivity. That's certainly possible, but commercially speaking, cloud adoption is growing and will slowly dominate most deployment scenarios.

On-premises security is critical, because it is the core of the company, and that's where the majority of the users will be accessing resources. When an organization decides to extend their on-premises infrastructure with a cloud provider to use IaaS (entry point 1), the company needs to evaluate the threats for this connection and the countermeasure for these threats through a risk assessment.

The last scenario description (entry point 4) might be intriguing for some skeptical analysts, mainly because they might not immediately see how this scenario has any correlation with the company's resources. Yes, this is a personal device with no direct connectivity with on-premise resources. However, if this device is compromised, the user could potentially compromise the company's data in the following situations:

• Opening a corporate email from this device
• Accessing corporate SaaS applications from this device
• If the user uses the same password [8] for his/her personal email and his/her corporate account, this could lead to account compromise through brute force or password guessing

Having technical security controls in place could help mitigate some of these threats against the end user. However, the main protection is continuous use of education via security awareness training.

The user is going to use their credentials to interact with applications in order to either consume data or write data to servers located in the cloud or on-premise. Everything in bold has a unique threat landscape that must be identified and treated. We will cover these areas in the sections that follow.

According to Verizon's 2017 Data Breach Investigations Report [9], the association between threat actor (or just actor), their motives, and their modus operandi vary according to the industry. However, the report states that stolen credentials are the preferred attack vector for financial motivation or organized crime. This data is very important, because it shows that threat actors are going after user's credentials, which leads to the conclusion that companies must focus specifically on authentication and authorization of users and their access rights.

The industry has agreed that a user's identity is the new perimeter. This requires security controls specifically designed to authenticate and authorize individuals based on their job and need for specific data within the network. Credential theft could be just the first step to enable cybercriminals to have access to your system. Having a valid user account in the network will enable them to move laterally (pivot), and at some point find the right opportunity to escalate privilege to a domain administrator account. For this reason, applying the old concept of defense in depth is still a good strategy to protect a user's identity, as shown in the following diagram:

Figure 2: Multi-layer protection for identity

In the previous diagram there are multiple layers of protection, starting with the regular security policy enforcement for accounts, which follow industry best practices such as strong password requirements, including frequent password changes and high password strength.

Another growing trend to protect user identities is to enforce MFA. One method that is seeing increased adoption is the callback feature, where the user initially authenticates using his/her credentials (username and password), and receives a call to enter their PIN. If both authentication factors succeed, they are authorized to access the system or network. We are going to explore this topic in greater detail in Chapter 7, Chasing a User's Identity. Another important layer is continuous monitoring, because at the end of the day, it doesn't matter having all layers of security controls if you are not actively monitoring your identity to understand the normal behavior, and identify suspicious activities. We will cover this in more detail in Chapter 12, Active Sensors.

Applications (we will call them apps from now on) are the entry point for the user to consume data and to transmit, process, or store information onto the system. Apps are evolving rapidly, and the adoption of SaaS-based apps is on the rise. However, there are inherited problems with this amalgamation of apps. Here are two key examples:

• Security: How secure are these apps that are being developed in-house and the ones that you are paying for as a service?
• Company-owned versus personal apps: Users will have their own set of apps on their own devices (BYOD scenario). How do these apps jeopardize the company's security posture, and can they lead to a potential data breach?

If you have a team of developers that are building apps in-house, measures should be taken to ensure that they are using a secure framework throughout the software development lifecycle, such as the Microsoft Security Development Lifecycle (SDL) [10]. If you are going to use a SaaS app, such as Office 365, you need to make sure you read the vendor's security and compliance policy [11]. The intent here is to see if the vendor and the SaaS app are able to meet your company's security and compliance requirements.

Another security challenge facing apps is how the company's data is handled among different apps, the ones used and approved by the company and the ones used by the end user (personal apps).

This problem becomes even more critical with SaaS, where users are consuming many apps that may not be secure. The traditional network security approach to support apps is not designed to protect data in SaaS apps, and worse, they don't give IT the visibility they need to know how employees are using them. This scenario is also called Shadow IT, and according to a survey conducted by Cloud Security Alliance (CSA) [12], only 8 percent of companies know the scope of Shadow IT within their organizations. You can't protect something you don't know you have, and this is a dangerous place to be.

According to Kaspersky Global IT Risk Report 2016 [13], 54 percent of businesses perceive that the main IT security threats are related to inappropriate sharing of data via mobile devices. It is necessary for IT to gain control of the apps and enforce security policies across devices (company-owned and BYOD). One of the key scenarios that you want to mitigate is the one described in the following diagram:

Figure 3: BYOD scenario with corporate app approval isolation

In this scenario, we have the user's personal tablet that has approved applications as well as personal apps. Without a platform that can integrate device management with application management, this company is exposed to a potential data leakage scenario.

In this case, if the user downloads the Excel spreadsheet onto his/her device, then uploads it to a personal Dropbox cloud storage and the spreadsheet contains the company's confidential information, the user has now created a data leak without the company's knowledge or the ability to secure it.

Data

We finished the previous section talking about data. It's always important to ensure that data is protected, regardless of its current state (in transit or at rest). There will be different threats according to the data's state. The following are some examples of potential threats and countermeasures:

These are only some examples of potential threats and suggested countermeasures. A deeper analysis must be performed to fully understand the data path according to the customer's needs. Each customer will have their own particularities regarding data path, compliance, rules, and regulations. It is critical to understand these requirements even before the project is started.

To analyze the cybersecurity challenges faced by companies nowadays, it is necessary to obtain tangible data, and evidence of what's currently happening in the market. Not all industries will have the same type of cybersecurity challenges, and for this reason we will enumerate the threats that are still the most prevalent across different industries. This seems to be the most appropriate approach for cybersecurity analysts that are not specialized in certain industries, but at some point in their career they might need to deal with a certain industry that they are not so familiar with.

Old techniques and broader results

According to Kaspersky Global IT Risk Report 2016 [14], the top causes for the most costly data breaches are based on old attacks that are evolving over time, which are in the following order:

• Viruses, malware, and Trojans
• Lack of diligence and untrained employees
• Phishing and social engineering
• Targeted attack
• Crypto and ransomware

Although the top three in this list are old suspects and very well-known attacks in the cybersecurity community, they are still succeeding, and for this reason they are still part of the current cybersecurity challenges. The real problem with the top three is that they are usually correlated to human error. As explained before, everything may start with a phishing email that uses social engineering to lead the employee to click on a link that may download a virus, malware, or Trojan.

The term targeted attack (or advanced persistent threat) is sometimes unclear to some individuals, but there are some key attributes that can help you identify when this type of attack is taking place. The first and most important attribute is that the attacker has a specific target in mind when he/she/they (sometimes they are sponsored groups) starts to create a plan of attack. During this initial phase, the attacker will spend a lot of time and resources to perform public reconnaissance to obtain the necessary information to carry out the attack. The motivation behind this attack is usually data exfiltration, in other words, stealing data. Another attribute for this type of attack is the longevity, or the amount of time that they maintain persistent access to the target's network. The intent is to continue moving laterally across the network, compromising different systems until the goal is reached.

One of the greatest challenges in this area is to identify the attacker once they are already inside the network. The traditional detection systems such as intrusion detection systems (IDS) may not be enough to alert on suspicious activity taking place, especially when the traffic is encrypted. Many researchers already pointed out that it can take up to 229 days between infiltration and detection [15]. Reducing this gap is definitely one of the greatest challenges for cybersecurity professionals.

Crypto and ransomware are emerging and growing threats that are creating a whole new level of challenge for organizations and cybersecurity professionals. In May 2017, the world was shocked by the biggest ransomware attack in history, called WannaCry. This ransomware exploited a known Windows SMBv1 vulnerability that had a patch released in March 2017 (59 days prior to the attack) via the MS17-010 [16] bulletin. The attackers used an exploit called EternalBlue that was released in April 2017, by a hacking group called The Shadow Brokers. According to MalwareTech [18], this ransomware infected more than 400,000 machines across the globe, which is a gigantic number, never seen before in this type of attack. One lesson learned from this attack was that companies across the world are still failing to implement an effective vulnerability management program, which is something we will cover in more detail in Chapter 16, Vulnerability Management.

It is very important to mention that phishing emails are still the number one delivery vehicle for ransomware, which means that we are going back to the same cycle again; educate the user to reduce the likelihood of successful exploitation of the human factor via social engineering, and have tight technical security controls in place to protect and detect.

The shift in the threat landscape

In 2016, a new wave of attacks also gained mainstream visibility, when CrowdStrike reported that it had identified two separate Russian intelligence-affiliated adversaries present in the United States Democratic National Committee (DNC) network [19].

According to their report, they found evidence that two Russian hacking groups were in the DNC network: Cozy Bear (also classified as APT29) and Fancy Bear (APT28). Cozy Bear was not a new actor in this type of attack, since evidence has shown that in 2015 [20] they were behind the attack against the Pentagon email system via spear phishing attacks.

This type of scenario is called a Government-sponsored or state-sponsored cyber-attack, but some specialists prefer to be more general and call it data as a weapon, since the intent is to steal information that can be used against the hacked party.

The private sector should not ignore these signs. According to a report released by the Carnegie Endowment for International Peace, financial institutions are becoming the main target for state-sponsored attack. In February 2019 multiple credit unions in the United States were targets of a spear-phishing campaign, where emails were sent to compliance officers in these credit unions with a PDF (which came back clean when ran through VirusTotal at that time), but the body of the email contained a link to a malicious website. Although the threat actor is still unknown, there are speculations that this was just another state-sponsored attack. It is important to mention that the US is not the only target; the entire global financial sector is at risk. In March 2019 the Ursnif malware hit Japanese banks. Palo Alto released a detailed analysis of the Ursnif infection vector in Japan, which can be summarized in two major phases:

1. The victim receives a phishing email with an attachment. Once the user opens up the email, the system gets infected with Shiotob (also known as Bebloh or URLZone).

2. Once in the system, Shiotob starts the communication with the command and control (C2) using HTTPS. From that point on, it will keep receiving new commands.

For this reason, it is so important to ensure that you have continuous security monitoring that is able to leverage at least the three methods shown in the following diagram:

Figure 4: Continuous security monitoring, facilitated by traditional alert systems, behavioral analysis, and machine learning

This is just one of the reasons that it is becoming foundational that organizations start to invest more in threat intelligence, machine learning, and analytics to protect their assets. We will cover this in more detail in Chapter 13, Threat Intelligence. Having said that, let's also realize that detection is only one piece of the puzzle; you need to be diligent and ensure that your organization is secure by default, in other words, that you've done your homework and protect your assets, trained your people and continuously enhance your security posture.