A comprehensive data strategy is essential for any organisation in the digital era. At the Data 2030 Summit this week, Mohamed Ashraf Ghazala, Lead Data Architect at Banque Du Caire, will present his expertise on establishing a comprehensive data strategy and proper data architecture to leverage the full potential of data and sustain the digital future of the company.
Mohamed Ghazala has more than 12 years with deep experience in modern data management solutions, supporting data-driven transformation projects, smart technology solutions and modern architecture using agile methodologies. Mohamed is currently leading Data Architecture and Governance in one of the biggest Egyptian banks, participating in a strategist role by defining effective data strategy aligning business strategy to future digital transformation and building data analytics platforms.
Hyperight: Hi Mohamed, we are very excited to welcome you as a speaker at the Data 2030 Summit. Your session will be focusing on Unlock the robust of effective data strategy and data cloud transformations. As you explain in your synopsis, Data Value has become an important strategic goal to fasten decision making. So having comprehensive data strategy is crucial to align it with the business strategy. What should a comprehensive data strategy look like?
Mohamed Ashraf Ghazala: Hi Ivana, I am so grateful to be joining Data 2030 Summit as a speaker and share my thoughts with all attendees. A comprehensive data strategy should first to be aligned with the organization’s business strategy. As we all know, all the organizations are establishing their digital transformation strategy in order to be aligned with the future revolution of technology in all the industrial domains. And Data is the core pillar of any digital strategy so it is very important to be derived with this organization strategy with defining vision and missions in a reliable road map.
Start with people culture and mindsets, measure the capabilities of Data management in your organization, define risk and challenges and how mitigation can help drive the organization to reach the real goal of data value. Modern data management methodologies and proper data architecture should take place. Understand the data and it’s metadata. Integrate all the data in one single point of truth. Provision the data and make it easily accessible to the whole employees to increase analytics and fasten the decision making. Empower technology of Machine Learning and AI to transform data from descriptive analytics to predictive analytics and prescriptive analysis.
Hyperight: One of the main points in your talk is cloud migration. What according to you are the common issues to avoid when migrating to cloud, and most important considerations for successful cloud migration?
Mohamed Ashraf Ghazala: Nowadays, we are in the era of a massive collection of data from tremendous sources and types of data: internal transactions, social media, mobile devices and automated sensors. This is where the benefit of Big Data Analytics takes place for generating and finding useful information. And there are multiple reasons for organizations to switch to cloud in order to gain better performance, costs, flexibility, scalability, agility, and maintenance of digital transformation. When we talk about issues and problems of cloud migrations many organization faces, we found most of these organizations are afraid of the network or services outages. Downtime is one of the most critical and common issues that might be faced in cloud migration. However, backup environments and strategies play a crucial role to avoid such challenges. Another very important issue companies faces are Data privacy and security as the internet security is the biggest issue IT company are afraid of. Cyberattacks are growing rapidly and data breaches are increasing. So all possible security solutions have to be developed to ensure no data leakage or breach. Interoperability is another demanding issue in cloud migration as it refers to the ability of systems to communicate to each other and it is not easy for an organization to establish proper communication between their applications so it is recommended to adopt internal processes with the cloud provider framework to achieve the proper success of communications.
Hyperight: Hybrid-cloud and multi-cloud are two strategies that are quite popular for enterprises creating their cloud strategies. They are often confused terms. So could we differentiate between them and explain their benefits, but also challenges?
Mohamed Ashraf Ghazala: Both Hybrid cloud and multi-cloud refer to cloud deployments that integrate more than one cloud platform. But the difference is in the cloud infrastructure in which is included. A hybrid cloud infrastructure merger two different types of clouds, for example, blends public and private clouds. However, multi-cloud is combining different services from the same type of cloud, especially in the public cloud. Organizations may also adopt more to hybrid clouds deployments to keep some of their sensitive data and processes controlled through their environments private clouds or on their own premises while taking the advantages of resources and infrastructures on public clouds. Still, the challenges remain the same for complexity, costs, downtime, privacy, security, operational efficiency, and latency.
Hyperight: Data governance is a key component of any data strategy. As companies increasingly move data to the cloud, designing a data governance strategy for a cloud-first world becomes imperative. How should businesses approach it?
Mohamed Ashraf Ghazala: A Data Governance framework is key enablement of any data strategy. When it comes to transforming data into cloud, Data Governance is a critical concern as organizations are allowing hosting their enterprise data at risk by leveraging the data to travel to a far-away location. Data should be protected and governed as it is the company’s assets. Cloud data governance goes on an additional dimension of complexity in multi-cloud or hybrid cloud computing environments, where data is found in multiple places and data governance protocols (authorizations, policies, metadata, data dictionaries, etc.) are inconsistent amongst data sources. It is recommended to standardize the policies of data access, well understand all the data elements with enhancing the organization metadata, measure the data quality and apply DQ rules to have more high quality of data, comply with regulations such as GDPR. IT is very important to ensure Data security, protection, retention policies to the cloud data.
Hyperight: As a final point, what are the trends in data management we can expect in the next couple of years?
Mohamed Ashraf Ghazala: As we are in the era of Big Data, the amount of data is growing rapidly Big Data address forward many challenges due to its several characteristics like volume, velocity, variety, variability, value and complexity. Big Data concern large-volume, complex, growing data sets with multiple, autonomous sources. Big Data are now rapidly expanding in all science and engineering domains, including physical, biological and biomedical sciences. The exponential growth of Data is becoming tremendous increasing every second. Such large sets are challenging for storing and processing the data. It has been solved by open-sources ecosystems such as Hadoop and NoSQL. However, these open sources required high capabilities of data engineering for manual configurations and troubleshooting. Also, business is starting to migrate big data to the cloud, either hybrid clouds or multi-clouds.
Machine learning is continuing to change the role of technology in Data management, it is becoming more sophisticated for potential uses. Fraud detections, retail trend analysis, robotics, & chatbots are becoming the future trends of ML. Many algorithms were defined earlier in the analysis of large datasets. There are many types of classification algorithms such as tree-based algorithms (C4.5 decision tree, bagging and boosting decision tree, decision stump, boosted stump, and random forest), neural-network, Support Vector Machine (SVM), rule-based algorithms (conjunctive rule, RIPPER, PART, and PRISM), naive Bayes, logistic regression. It is all under the umbrella of Soft Computing and Data Science. Data Scientist and CDOs will be in high demand in the labour market.