Shifting to the Cloud: the story of a French biotech startup
New molecule discovery and development is a highly complex activity generating a large amount of information and data. This data is becoming more complex with time. Each new scientific technique and method developed in the laboratory creates new challenges for IT to address.
Fortunately, new research and discovery informatic solutions arise to help support scientists and companies to fully benefit from these emerging developments.
Nonetheless, in this cutting-edge environment, some standard solutions readily accepted by other industries are often dismissed because of preconceived ideas. The Cloud is a standard that needs to be embraced.
Despite a few downsides, switching from an on-premises (local) to a Cloud environment can bring a lot of benefits, to research-driven organizations.
This article will outline an example of a French Biotech Startup that migrated its scientific software applications toolset to the Cloud and reveal the real-life advantages of this migration.
A little bit of history
The company develops and optimizes enzymes by generating libraries of variants. It then identifies those with the best activity using a high-throughput screening process.
The software suite used to support their activity is composed of a sample management solution (Discngine Sample for Assay), and a plate-based assay analysis solution (Discngine Assay) including an algorithm to parse automatically chromatograms. The applications were initially installed in 2014 and are used almost daily, generating 3TB of historical data.
The desire to migrate their scientific applications to a Cloud environment was not an isolated wish from a part of the organization, but a larger company-wide initiative to migrate their entire Information System.
This decision was made a few months before the beginning of the COVID-19 pandemic and revealed itself to be a great strategic move.
Shifting in 3 simple steps
Beyond management's desire, other criteria were considered when Discngine and its customer explored their options:
The termination of the local server hosting service by the historical provider
An obsolete local server
The cost evaluation of acquiring and maintaining new equipment.
The lack of internal knowledge and skills to manage and maintain the infrastructure (database, servers, applications, etc..)
All these reasons combined led to choosing a Cloud shift as a solution.
Moreover, working hand in hand with its customer, Discngine designed a tailored solution and strategy to make this migration happen. This Cloud shift strategy followed a 3-step process over 5 months (during the pandemic).
First, to meet performance and cost expectations, as well as to overcome the lack of in-house expertise, Discngine selected Oracle as the infrastructure provider. This allowed its customer to benefit from a new solution called Autonomous DataBase which automatically updates and maintains the environment.
Some technical details
Regularly updating software applications is important to benefit from the most advanced security settings and technical support.
Here are some technical details about this cloud shift project:
Infrastructure and software updates :
-
OCI = Oracle Cloud infrastructure (ATP = Autonomous Transaction Processing).
-
From Oracle 10 to Oracle 19
-
From TIBCO Spotfire 7.6 to 10.2 with Connector upgrade
-
From APEX 4.1 to 20.2
Data migration strategy
Approximately 3TB of historical data needed to be migrated. The Data export was done with Data Pump and automatically transferred with OCI using multithreaded data transfers to optimize performance. Foreign keys were verified, and some random value checks were manually performed by Discngine & the customer.
-
Data from 2014 to 2019: 2.5 TB
-
Data from 01/2020 to 06/2020: 500 GB
-
Data from 06/2020 to 09/2020: few MB
Second, to ensure component compatibility and to avoid any loss of functionality, Discngine progressively migrated the versions of the necessary framework and the applications. Some improvements were also performed on the applications at this stage because these old versions of the software suite had been specifically modified to fit the scientific needs of the company.
Third, to minimize operational downtime, all research data was migrated in multiple steps. Much of the historical data was migrated to the new environment first so that only a small amount of data migration was required when moving applications to the new environment. This strategy allowed some users to test the new environment with real data while the company continued to operate on its local server.
In the end, only one weekend of service interruption was necessary to migrate the data leftover and switch users to the Cloud.
In the course of this project, one particular challenge was faced, which was to connect an on-prem barcode printer to this new cloud environment.
But after replacing the local network communication protocol for devices with a new adapted secured protocol, this was made possible, and the initial printer was able to operate with the new Cloud environment.
Exceeding expectations
The migration to the new Cloud environment revealed several benefits, some of which were expected, and others turned out to be a pleasant surprise.
1. Automatic maintenance and improved security.
Indeed, the new service Autonomous DB provided by Oracle includes the autonomous management of the environment, combining the agility of the cloud and the responsiveness of adaptive machine learning. The company can be certain that they will always benefit from the latest versions of their Cloud environment. In terms of security, the measures of the provider (ISO27001) offer them a high level of security, that would be difficult to match if they had their servers on-site. Moreover, the database, in tripled-mirror configuration, recovers automatically from failures, and nightly backups are stored in different locations, ensuring they will never lose all their research work.
2. Performance
The time required to perform regular operations in the new environment has been reduced tremendously. This benefit completely exceeded their expectation. Some numbers to illustrate:
Sample research which usually took minutes is now done in less than 10 seconds.
The Analysis of a run improved from a few hours (sometimes left as an overnight task) to only a few minutes (15 minutes max.), including the graphical analysis of chromatograms.
Indeed, by migrating to the Cloud, the French company upgraded its software as well as the hardware used in their activity. Great hardware (SSD, CPU, GPU,…) are available on servers and is the responsibility of the service provider.
3. Cost reduction
With their servers being online only when required by their research activity, the company is saving an important part of running costs as compared to having a server always running.
4. Remote work facilitated
This advantage may seem logical with the global pandemic we are currently experiencing, but their choice was made before the beginning of COVID. Being able to easily shift to a "work from home" model was another great unexpected benefit from working using a Cloud environment.
“Don’t be afraid of the Cloud!
It can really bring great value to your business and research.”
Shifting to the Cloud could be your next best decision
The experience of this French Biotech Startup is a nice example of the value brought by using a Cloud environment.
The flexibility experienced managing hours of operation, accessing the solutions, and getting automatic updates, as well as their performance gains, are all due to their migration to a Cloud environment.
Since the first Cloud offerings (not suitable or attractive to research organizations), Cloud offers now have experienced a leap forward in performance, quality and security.
These new offers can be suitable for companies of all sizes and especially for companies with rapidly evolving needs.
Finally, if operating such a change is challenging for you, this use case also demonstrates that bringing another party to help with the project can also be a great idea to guide you in this dematerialized world.