Pharmaceutical giant AstraZeneca, along with its research partner from the University of Oxford, has been at the forefront of the global effort to rapidly research, develop and deploy a vaccine that could help slow the spread of the novel Covid-19 coronavirus throughout the world. world.
The first of 100 million doses of the Oxford-AstraZeneca vaccine that the UK government has procured was administered in early January 2021 and, at the time of writing, GP surgeries across the country they are beginning to receive their reservations.
While much of the media attention that the Oxford-AstraZeneca vaccine has attracted to date has focused on the rapid pace of its development, much of the behind-the-scenes work that has made it possible to confer large-scale protection against Covid-19 within the UK population has relied on cloud computing.
AstraZeneca’s director of global infrastructure services Scott Hunter is responsible for the pharmaceutical company’s cloud platforms and innovation solutions for the infrastructure and cybersecurity portion of its business.
The company has four of the major public cloud provider platforms to carry out its work, including Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform, and Alibaba.
It also relies on the interconnection capabilities of colocation giant Equinix to make these four companies’ respective cloud technologies available through AstraZeneca’s own data centers.
“We have put together our own methods of operations on top of that, but at the same time we take advantage of some of the niche capabilities like natural language processing (NLP) with search in Azure, and we use a lot of infrastructure as a Service. [IaaS] on AWS for research and development, ”says Hunter.
Hunter’s 112-person team has responsibility for architecture, design and governance, and controls what the company does with its hybrid multicloud environment, known as the AZ Cloud.
Data ingestion processes
With regard to AstraZeneca’s Covid-19 vaccine development and production efforts, the biggest challenge it has faced is lack of time.
Before Covid, antibodies could only be tested in a small subset of the population and gradually increased over a period of 12 to 24 months or more. However, with limited time, AstraZeneca needed to assimilate adverse event data from two billion doses that were rolled out at the same time around the world.
Scott Hunter, AstraZeneca
“It is incredibly important to collect data, particularly adverse event data, to ensure that it is fed back to relevant organizations in real time. Part of that process allows us to cross-reference this data with existing data from our own patients to ensure that there is no adversity with our patients and the therapies they are currently taking, ”says Hunter.
To do this, the Hunter team has worked closely with AstraZeneca’s connectivity teams to ensure that company data is used where it is needed, thanks to an abstraction layer provided by storage giant NetApp. .
However, this was not always the way the company worked with its data platforms. About five years ago, the pharmaceutical company predominantly used Hitachi Data Systems (HDS) and Hewlett Packard Enterprise (HPE), and NetApp was used for its file-based solutions. Decided to review this and made a Request for Information (RFI).
“We found that in addition to offering capabilities similar to what AstraZeneca was getting with HDS and HPE, the differences in how the company could use NetApp in the cloud were compelling, and from a price point per terabyte it was better, so the Workload optimization and consolidation with a single vendor and getting a good deal made sense to us, ”says Hunter.
NetApp is being used to provide a data structure that enables AstraZeneca to collect data from the four cloud providers it works with, federate this data with partners and research institutes, and inform the development of other Covid-19 treatments and therapies.
“The important element of the data fabric used for the hybrid multicloud was the movement of workloads from a private cloud to a public cloud; the benefits for scientists is that they know that the data will be available all the time, ”adds Hunter.
“The biggest challenge before the hybrid multi-cloud approach was data conversion. The data structure allows us to run NetApp efficiently in the cloud, so there is no conversion to do; we can move data between the local and public cloud, and start the service in less than 10 minutes ”.
Vaccine deployment underway
Now that the vaccine is spreading to priority groups across the UK, and other territories around the world as well, the benefits of having access to a global hybrid cloud environment based on multi-vendor technologies are really coming to the fore.
“We host a large number of workloads and services in the public cloud because, as you can imagine, deploying a vaccine around the world in countries that do not have a data center location or are not close to one of our key data centers, it makes sense that you leverage any of the four public clouds to ensure you have the best range of services so we can get a common approach to learning what to do, ”says Hunter.
These lessons focus on developing new patient advice, governance, details on adverse effects, and follow-up, based on feedback the company receives.
Additionally, AstraZeneca relies on tools such as NLP to ensure that it can create websites with vaccine information in different languages. Normally, that information would be made available to patients in the form of a physical brochure, but since time is of the essence, that information must now be transmitted online.
The company has implemented a security-by-design approach, which Hunter suggests has been essential, particularly at this time when R&D data and information is a huge source of attraction for cybercriminals.
“As you can imagine, right now AstraZeneca is at the forefront and is the main thought of many bad actors: we have five to six million events per day, so safety is key for us,” he says. “If you think about it from a data protection or disaster recovery point, the data structure ensures that we don’t lose any data inappropriately.”