Case Study Summary/Organizational Need
Techfite is a Texas-based company that specializes in medical technology. Its proximity to the Johnson Space Center has determined the organization’s direction of partnership in that it works closely with NASA. More specifically, Techfite develops the solutions that allow astronauts to spend more time in space without consequences for their health. In light of the type of the job, Techfite has been forced to comply with major governmental requirements under the circumstances when budgets are being cut. In this regard, the company’s desire to explore additional partnerships overseas is reasonable, but precautions should be taken to ensure the integrity of cooperation.
Currently, there are serious restrictions in terms of data handling that are observed by Techfite’s competitors, especially in the European Union. At the same time, Techfite looks to launch new partnerships with the European Space Agency, the Canadian Space Agency, and the Japanese Space Agency. Undoubtedly, the company’s expertise that originates from its NASA projects creates a favorable foundation upon which new agreements can be built. However, a certain reorganization of the data-handling architecture appears necessary for the envisaged development. First of all, the expansion of Techfite’s network is likely to prompt the creation of new subsidiaries across national borders and continents. Thus, the company will have a higher need for prolonged and accessible data storage. Second, in addition to being stored for a longer period, the data will exhibit a stronger need for security. The expansion of the framework will create new potential points of security breach.
Ultimately, Techfite needs a system of data storage, in which the information is accessible for all authorized subsidiaries. Simultaneously, malicious activity needs to be prevented and eliminated, considering the nature of the data handled. Currently, scans 10,000 log files manually per week with a 30% success rate in identifying intrusion signature patterns and has only 4 terabytes of storage allocated locally for log storage. This report outlines a new solution that can be used to enhance Techfite’s capability under the envisaged circumstances.
Proposed Emerging Technology Solution
In the 21st century, deep learning has become one of the leading emerging technologies with a broad spectrum of application. According to Bashar (2019), this technology’s primary feature consists of its replication of the human-like learning mechanisms, allowing the system to adjust to needs of the user. Deep learning allows for the classification of tasks that are supplied in a variety of forms, including text, images, and sounds. Guo et al. (2018) note that the initial lack of transparency has limited the range of scenarios in which deep learning is applied. However, current studies in this area highlight the increased potential of deep learning for the identification of malicious activity within cloud networks.
Such architecture is strongly required by Techfite’s expansion plans, as clouds enable simultaneous access to the database by all subsidiaries. Furthermore, it will lift hardware restrictions on the amount of data stored, allowing Techfite to expand beyond the current 4 terabytes. Databricks (2021a) is a vendor that can provide the company with a bespoke solution that adapts the potential of deep learning for Techfite’s security needs. It relies on the predictive analytics solutions to identify abnormal behavior within the network, filtering out malicious activity with a higher precision rate than the current one (Lv, 2021). It is currently in the Visionary quadrant, inevitably making its way toward the Leader status.
Adoption Process
The transition toward deep learning-based data handling should be completed in several steps to allow for a gradual transformation. First, pilot tests should be executed with the contractor outside the actual Techfite framework. Second, once an understanding of the organization’s goals is reached, the network is to be applied to a select sector of data within the primary headquarters of the company. This stage is required to test the efficiency of the data-handling framework in an actual environment while mitigating the risks of the full-scale implementation. Third, when practical issues inevitably arise, necessary corrections will be made to the bespoke predictive algorithm. Finally, the framework can be expanded to all subsidiaries of Techfite’s expanded network with simultaneous personnel trainings. Table 1 reflects the adoption process in its alignment with the STREET paradigm.
Table 1. Adoption Process.
Technology Impact
The implementation of the deep learning-assisted data handling cloud will create a better system of data-handling. First of all, the information will be accessible by all authorized subsidiaries, including the potential overseas offices. It is important to keep the database up-to-date for all users, and clouds allow for constant synchronization. At the same time, deep learning algorithms will be taught to separate standard and abnormal actions of all users that operate in the cloud. This way, the precision of malicious entry identification is expected to increase from the current 30% to 80% (Databricks, 2021b). This way, the integrity of the framework will be maintained.
On the other hand, such digital transformations bear risks, as well. Databricks (2021b) acknowledge that the implementation of deep learning-based clouds for data-handling requires the use of costly infrastructure. In addition, personnel need to be provided some advanced training to master this sophisticated solution and use to its full potential. Hence, the current 4-person IT department will have to be expanded to at least 6 or 7 members, one of whom will be specialized in deep learning responsible for the new framework entirely.
Technology Comparison
Deep learning-based cloud framework is not the only viable option for Techfite’s needs. Table 2 introduces two alternatives in comparison with the proposed technology. As can be inferred, while deep learning is more complex and costly, it can yield better results in the long-term due to its increased growth potential.
Adoption Success
The success of the new technology’s adoption will be judged by two key metrics. Currently, Techfite manually checks 10,000 logs per week, whereas automation can yield an exponential increase in this regard (up to 100% coverage). Next, the key metric will be the success rate of the log verification. As of now, the success rate is 30%, whereas deep learning will increase it. Even though the range of 80%-100% seems attainable in the long-term, as much as 50%-60% success will suffice to state the success of the implementation.
Conclusion
Ultimately, Techfite faces a potential expansion with the inclusion of new subsidiary offices in its framework. In light of the importance of secure data-handling, a reliable, yet convenient framework needs to be introduced. Deep learning-based cloud solutions can meet the long-term needs of the organization, resulting in a higher number of logs checked with better success rates. This opportunity deserves additional attention, meaning that its exploration is highly recommended.
References
Bashar, A. (2019). Survey on evolving deep learning neural network architectures. Journal of Artificial Intelligence and Capsule Networks, 2(2), 73-82. Web.
Databricks. (2021a). Databricks. Web.
Databricks. (2021b). Solutions. Web.
Guo, W., Mu, D., Xu, J., Su, P., Wang, G., & Xing, X. (2018). LEMNA: Explaining deep learning based security applications. CCS ’18: Proceedings of the 2018 ACM SIGSAC Conference on Computer and Communications Security, 1(1), 364-379. Web.
Lv, Z., Qiao, L., Li, J., & Song, H. (2021). Deep-learning-enabled security issues in the Internet of Things. IEEE Internet of Things Journal, 8(12), 9531-9358. Web.
Mello, J. P. (2020). 5 emerging security technologies set to level the battlefield. TechBeacon. Web.