第14章サイバー資源の品質とサイバーコンピューティングのガバナンス

In our development pyramid, where cloud-centric platforms are at the top (layer 3), cloud and big data analytics form an overarching umbrella. The process of analyzing that data to find correlations and trends that can be used for decision-making is called big data analytics. Incorporating the Platform as a Service (PaaS) aspect for big data analytics provides a cloud-based development environment that can solve many development problems and provides methods to analyze and enhance system performance.

A big challenge with data analytics from a TRM perspective is time. It takes time to properly analyze data collected to make informed business decisions. Big data only helps information management when data is converted into information on time. While cloud providers do their best to offer 24/7 customer service, issues arise when third parties and containers are introduced. The data may not be immediately available from the third-party vendors, or they may not want to share their data. Having clear agreements with third parties about what data will be shared and what data will not be available is important.

A second challenge with big data solutions is collecting the wrong data. For instance, firewall log files might record time stamps of external IP address access attempts, but if the firewall doesn’t account for IPv4 to IPv6 conversions, some IP address data could be lost. If the firewall only handles IPv4 and the attacker uses IPv6, the log files may be incorrectly formatted, resulting in incorrect or incomplete data analysis.

A third challenge for data analysis and PaaS is bandwidth. Proper data analysis requires large amounts of computing power and hardware, slowing network communications during data transmission. Storage is another issue, as analyzing multiple days of data requires additional storage locations, often leading to cloud storage and virtualized solutions. Companies like Google generate revenue by doing big data analysis and selling their findings to companies for marketing strategies. Managing over 1.8 billion Google accounts with a “free” 10GB cloud storage drive necessitates virtualization and an extensive security plan.

A service-level agreement (SLA) must be established when customers enter into agreements with PaaS platforms. The SLA outlines the services that the platform provides. Companies can estimate growth and need for their organization using SLAs. However, SLAs are not covered in the TRM, and since they are not actual “ilities,” they may fail to include actionable items. This means that the SLAs are “what if” actions. For example, if there is a data breach, this will happen, making the SLA reactive rather than proactive. The TRM aims to prevent issues before they arise. SLAs are criticized for lacking assurance and privacy protections, as data breaches can put user information at risk. This highlights the importance of trust, which, while not directly mentioned in the TRM, is addressed in many different “ilities” in implied ways.

Trust in any service-providing system has a narrow margin for error. A trust relationship is established when consumers entrust their information to a company. To bolster this trust, companies should have information security policies detailing data security practices. However, exact methods should not be disclosed to prevent giving hackers helpful information. The trust relationship aims to make the consumer feel secure, and therefore increase their ease-of-use with the system. This trust can be a valuable metric for measuring quality based on customer satisfaction.

Earlier, we compared technology to a one-mile race, with IoT representing the last ten steps. Every technological advancement has led to the IoT, as consumers want more power and faster access to information. They are dissatisfied when companies fail to provide information quickly and efficiently. Cloud IoT, as a PaaS, provides more challenges than this small section of the book can cover. To start, IoT’s largest challenge is scalability. The rapid development of wearable devices, smart devices, smart cars, appliances, and artificial intelligence suggests no end to this growth. Consequently, IoT scalability is increasingly challenging to manage.

If scalability is out of control, then it is typically the case that security is out of control. Most TRM “ilities” encounter challenges with the IoT. Having a quality security plan for every possible third-party IoT device is not feasible and it is sufficient to analyze the performance of the underlying Bluetooth and Wi-Fi technologies. Developing better and stronger ways to secure communications between devices would be an excellent first step. There will always be an issue with hackers capturing data transmitted wirelessly.

The next challenge for IoT as a PaaS is storage. Where is all the captured data stored? Apple devices are stored in their iCloud, but other companies face security issues with stored data. Consider the data that is left behind on a device that is no longer in use. Best practices for disposing of computer hard drives and other storage devices don’t always apply to IoT devices and smartphones. We created regulations like the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) to address concerns arising from IoT storage situations.

Interoperability is a major challenge with IoT devices, given the vast types of devices and communication methods. Securing biometric data is a constant concern, and hacking into home monitoring systems with IoT devices poses significant safety risks. Since most IoT devices rely on Wi-Fi, improving interoperability efforts and security in wireless communications is necessary. Cloud-centric services are another concern, as all IoT devices are part of cloud technology. Developing interfaces for IoT devices is also challenging due to the need to accommodate cultural and language differences.