We shall be building Pakistan’s first truly digital lab; to reflect its true essence we are calling it “Digital Pakistan Lab”. The aim of the lab is to assist Pakistan’s transformation to digitization by providing a national platform of data analytics and cloud computing to serve multiple sectors and domains. We perceive the hardest problem in using data analytics in Pakistan in different sectors and domains is the user’s commitment to digitization since data analytics can only be performed on digital data.
The second challenge is the lack of customized data analytics platform for the problems our nation is encountering. Most of the data analytics applications require availability of High Performance Computing (HPC) platforms and large secure storage. In our country, there has also been a fear of losing critical data to adversaries. The proposed lab shall address all these challenges by adapting strategies that shall address all these issues. The lab proposes to develop a hardware software abstraction incorporating a broker architecture that would have the capability of incorporating all public funded HPC/clusters and data centers for effective use. The lab shall also build/port generic data analytic tools/algorithms to execute on these HPC platforms.
The lab also aims to develop application specific data analytic tools for problems pertaining to national needs. The framework shall provide access to researchers and developers f not only to use the developed hardware/software framework but also to augment their own developed algorithms and computing and storage resources in this national framework. For this purpose, Digital Pakistan Lab shall develop solutions for various applications in industrial, health, and energy sectors of the country in due course of the project to demonstrate Lab’s spectrum ranging from data analytics to high performance computing.
We envision seeing this lab becoming a true National Centre of Excellence in data analytics and cloud computing.
The goal of the Digital Pakistan Lab is to create a platform where multiple tools will be developed for multiple domains involving both commercial and non-commercial entities so that they can use these services effectively. Our scientific goal is to create innovation ecosystem based on applied R&D network of partners in government, industry, and academia. The hardest problem is the user’s commitment to digitize, lack of customized data analytics platform specific for national needs, ensuring security and secrecy of data. The lab will work to overcome these issues in Pakistan and will set a course for other initiatives as well to achieve the objectives of Vision 2025. The architecture will be realized by implementing a broker architectural pattern for distributed application. The purpose of broker architecture is to build a system as a set of decoupled interacting components to achieve flexibility, location transparency, maintainability, and scalability.
We propose to implement broker pattern in a way that clients see a single logical system which physically runs on various geographically separated platforms. Clients are registered and can use hardware or software services. Hardware services include high performance computing and storage resources through broker component. Broker component selects the target hardware platform whenever a client requests for a computing resource and forwards client’s request to that selected platform. Similarly, broker decides location of the physical storage on a client’s storage request. In order to use software services, clients can use any of the following three options
Web Services: Logic runs on clusters, data uploaded by clients to get their results
Dynamic Link Libraries (DLLs): Logic is downloaded by clients and they can embed that in their own source codes
Mobile Agents: Executable agents execute on client computers without any data migration to perform data analytics
This would allow the users across domains to securely use the published tools without any fear of data loss to adversaries. Lab shall provide data analytics and computing services across three domains:
Three types services will be provided:
HPC and storage resources through a broker architecture
Generic data analytics algorithms utilizing HPC
Domain specific data analytic algorithms
Any HPC and storage resources in public domain can be made part of the framework for effective utilization. This feature is extremely important since public funded clusters often remain offline or are not being optimally utilized.
Our core R&D objectives will be:
Analytics: The lab will turn numbers into knowledge that informs and empowers.
Decision sciences: In the digital economy, data with the science of decision making needs to work together.
Software and computational systems: Develop trustworthy software to help make sense of data and enable safe sharing.
Engineering and user experience design: To building novel technology through design process involving high performance computations thus making impact through innovation
The lab has a good business plan to ensure its sustainably beyond three years of funding. Right from its inception the lab shall be exploring opportunities to realize it business potential.
The lab shall sign 3-4 industries for their transformation to Industry 4.0 standard based on the developed technologies in the lab. The lab will offer to license data analytics algorithms for key decisions relating to production planning, inventory management and shop floor automation.
The lab shall create a large repository of automated ECG and intra-cardiac signals along with imagery related to ophthalmology including fundus and OCT images. The lab shall charge a fee for researchers across the globe to use this dataset.
The lab shall also offer arrhythmia detection, EMR based ophthalmology data analytics tools for licensing.
Lab shall sign 2-3 eye hospitals/ ophthalmology clinics for EMR and data analytics implementation.