5 more retail cybersecurity practices to keep your data safe beyond the holidays

This is the second article in a two-part series about retail cybersecurity during the holidays. Read part one for the full list of recommendations.

The holiday shopping season offers myriad opportunities for threat actors to exploit human nature and piggyback on the rush to buy and sell products in massive quantities online. Our previous post covered some network security basics for retailers. Let’s take a closer look at how retailers can properly configure and monitor their networks to help mitigate cyberattacks and provide customers with a safe shopping experience during the holiday season.

Baselining is the process of measuring normal amounts of traffic over a period of days or even weeks to discern any suspicious traffic peaks or patterns that could reveal an evolving attack.

Network traffic measurements should be taken during regular business hours as well as after hours to cover the organization’s varying activity phases. As long as the initial baseline is taken during a period when traffic is normal, the data can be considered reliable. An intrusion detection system (IDS) or intrusion prevention system (IPS) can then assist with detecting abnormal traffic volumes — for example, when an intruder is exfiltrating large amounts of data when offices are closed.

Below are some factors to consider when performing a baseline measurement that could be helpful in detecting anomalies:

• Baseline traffic on a regular basis.
• Look for atypical traffic during both regular and irregular times (e.g., after hours).
• Set alarms on an IDS/IPS for high and low thresholds to automate this process. Writing signatures specific to your company’s needs is a key element to an IDS/IPS working effectively and should be carried out by trained security specialists to avoid false alarms.
• Investigate any discrepancies upon initial discovery and adjust thresholds accordingly.
• Consider using an endpoint detection and response (EDR) solution to help security teams better identify threats, and to allow operations teams to remediate endpoints quickly and at scale.

A key preventative measure for retailers with a more mature security posture is running a penetration test. Simply put, the organization’s security team can allow a white hat hacker, or penetration tester, to manually try to compromise assets using the same tactics, techniques and procedures (TTPs) as criminal attackers. This is done to ascertain whether protections applied by the organization are indeed working as planned and to find any unknown vulnerabilities that could enable a criminal to compromise a high-value asset.

Manual testing should be performed in addition to automated scanning. Whereas automated tools can find known vulnerabilities, manual testing finds the unknown vulnerabilities that tools alone cannot find. Manual testing also targets the systems, pieces of information and vulnerabilities most appealing to an attacker, and specifically focuses on attempting to exploit not just technical vulnerabilities within a system, but business logic errors and other functionality that, when used improperly, can grant unintended access and/or expose sensitive data.

The key to a penetration test is to begin by assessing vulnerabilities and addressing as many of them as possible prior to the test. Then, after controls are in place, decide on the type of test to carry out. Will it be a black box test, where the testers receive no information about the target’s code and schematics? Or will it be a white box test, where organizations fully disclose information about the target to give the tester full knowledge of how the system or application is intended to work? Will it be in a very specific scope and only include customer-facing applications?

It can be helpful to scope a penetration test by taking the following three steps prior to launching the testing period:

1. Establish goals for the testing. Since penetration testing is intended to simulate a real-world attack, consider scenarios that are relevant to your organization. Giving thought to what type of data is at risk or what type of attacker you’re trying to simulate will allow the testers to more closely approximate threats relevant to your organization.
2. Draft a thorough contract to state the expectations and scope of the project. For example, if there are specific areas a penetration tester should not access based on criticality or sensitivity, such as production servers or credit card data, outline these points in the contract. Also, define whether the penetration testers should attempt to compromise both physical access and remote access to compromise networks, or if just one is preferred. Consider if you wish to have social engineering included within the test as well.
3. Have the vendor and its employees sign nondisclosure agreements (NDAs) to keep their findings confidential and ensure their exclusive use by the organization.

Penetration testers from reputable companies are thoroughly vetted before being allowed to conduct these tests. The retail industry can benefit from this type of testing because it mimics the actions of a threat actor and can reveal specific weaknesses about an organization. It can even uncover deficiencies in staff training and operational procedures if social engineering is included within the scope of the testing.

Log data collected by different systems throughout an organization is critical in investigating and responding to attacks. Bad actors know this and, if they manage to breach an organization and gain elevated privileges, will work to cover up their tracks by tampering with logs.

According to IBM X-Force Incident Response and Intelligence Services (IRIS) research, one of the most common tactics malicious actors employ is post-intrusion log manipulation. In looking to keep their actions concealed, attackers will attempt to manipulate or delete entries, or inject fake entries, from log files. Compromising the integrity of security logs can delay defenders’ efforts to find out about malicious activity. Additional controls and log monitoring can help security teams avoid this situation.

Below are some helpful tips and examples of security logs that must be checked to determine whether anything is out of the ordinary.

• Are your logs being tampered with? Look for altered timestamps, missing entries, additional or duplicate entries, and anomalous login attempts.
• Transfer old log files to a restricted zone on your network. This can help preserve the data and create space for logs being generated overnight.
• Use a security information and event management (SIEM) tool to assist with analyzing logs and identifying anomalies reported by your organization’s security controls.
• To include as many sources of information as possible, plug in endpoint, server, network, transaction and security logs for analysis by a SIEM system. Look for red flags such as multiple failed logins, denied access to sensitive areas, ping sweeps, etc.

Knowing which logs to investigate is also critical to successful log analysis. For example, point-of-sale (POS) systems are often installed on Microsoft Windows or Linux systems. It is therefore critical to review operating system logs for these particular endpoints. When it comes to POS networks, where many of the devices are decentralized, daily usage, security and application logs are good places to look for anomalies.

For network security, use logs from network appliances to determine failed or excessive login attempts, increases or decreases in traffic flow, and unauthorized access by users with inadequate privilege levels.

According to the National Retail Federation, online sales from November and December 2017 generated more than USD 138.4 billion, topping 2016 sales by 11.5 percent. This year is likely going to set its own record. With internet traffic volumes expected to be at their highest, online retailers that are unprepared could see the loss of sales and damaged reputation in the aftermath of the holiday season.

But preparing for extra shoppers is the least of retailers’ worries; attackers may take advantage of the festive time of year to extort money by launching distributed denial-of-service (DDoS) attacks against retail websites. These attacks work by flooding a website or network with more traffic than it can handle, causing it to cease accepting requests and stop responding.

To stay ahead of such attacks, online retailers can opt to use designated controls such as load balancers. Load balancers are an integral part of preventing DDoS attacks, which can affect POS systems storewide. With a well-coordinated DDoS attack, a malicious actor could shut down large parts of their target’s networks.

One best practice is to prepare before traffic peaks. Below are some additional tips for a more balanced holiday season.

• Preventing a DDoS attack can be an imposing undertaking, but with a load balancing device, most of this work can be automated.
• Load balancers can be either hardware devices or virtual balancers that work to distribute traffic as efficiently as possible and route it to the server or node that can best serve the customer at that given moment. In cases of high traffic, it may take several load balancers to do the work, so evaluate and balance accordingly.
• Load balancers can be programmed to direct traffic to servers dedicated to customer-facing traffic. Using them can also enable you to move traffic to the proper location instead of inadvertently allowing access to forbidden areas.

Load balancers are typically employed by larger companies with a prominent web footprint. However, smaller companies should still consider employing them because they serve a multitude of purposes. Keeping the load on your servers balanced can help network and website activity run smoothly year-round and prevent DDoS attacks from doing serious damage to your organization’s operations or web presence.

An incident response (IR) plan is essential to identifying and recovering from a security incident. Security incidents should be investigated until they have been classified as true or false positives. The more timely and coordinated an organization’s response is to an incident, the faster it can limit and manage the impact. A solid IR plan can help contain an incident rapidly and result in better protection of customer data, reduction of breach costs and preservation of the organization’s reputation.

If your enterprise does not have an IR plan, now is the time to create one. In the event that your enterprise already has a plan, take the time to get key stakeholders together to review it and ensure it is up-to-date. Most importantly, test and drill the plan and document its effectiveness so you’re prepared for the attack scenarios most relevant to your organization.

When evaluating an IR plan, consider the following tips to help accelerate your organization’s response time:

• Threat actors who compromise retail cybersecurity will typically turn stolen data around quickly for a profit on the dark web. Use dark web search tools to look for customer data that may have been compromised. Sometimes, data can be identified by the vendor that lost it, leading to the detection of an ongoing attack.
• Before an attack occurs, establish a dedicated IR team with members from different departments in the organization.
• Make sure each team member knows his or her precise role in the case of an incident.
• Keep escalation charts and runbooks readily available to responders, and make sure copies are available offline and duplicated in different physical locations.
• Test your IR strategy under pressure in an immersive cyberattack simulation to find out where the team is strong and what may still need some fine-tuning.

Increased vigilance is important for retailers during the holiday season, but these network security basics and practices can, and should, be maintained throughout the year. Remember, attackers don’t just wait until the holiday season to strike. With year-round preparation, security teams can mitigate the majority of threats that come their way.