5G, Multi-cloud, Kubernetes… Patrick McFadin shares his predictions for the year 2020 in order to help IT professionals, and especially developers, to approach this new year.
Thanks to the evolution of technologies, developers can now implement their applications on more specialized databases. We are moving from historical relational databases to new databases which lend themselves more to use cases such as graphs or time series data. Organizations should soon take more and more interest in so-called NoSQL databases in order to cope with the explosion in data volumes.
Many of these new projects will be based on cloud databases (DBaaS or Database-as-a-service) which will facilitate deployment and management. Some expertise in the cloud and in data management will always be necessary to understand the constraints (latency, scalability, resilience, costs) in order to determine the most suitable platform.
In 2019, an upsurge in production incidents due to poor configurations (security, deployment) was observed. Unless industry players facilitate the adoption of solutions designed to be more robust or all unsecured deployments are identified (laughter), this phenomenon is not going to run out of steam in 2020.
To remedy this, database editors and cloud providers who offer a managed service will have to put in place good security practices and meet all the conditions to impose their implementation. We must capitalize on the expertise around databases to develop better services – whether through the products themselves or their cloud version
Graph-oriented databases have made great strides lately. Knowledge graphs (eg UDF) and property graphs make it easier to model and analyze relationships between different objects.
As a growing number of organizations work to identify the existing relationships between objects, locations, and individuals, graph databases are expected to gain in popularity.
The growth of the graph in 2020 could however be slowed down by the lack of understanding surrounding this technology. This is explained by the scarcity of profiles with the specific skills required. As new solutions emerge to market, they should facilitate adoption, but those who do take the plunge can expect to have trouble starting up configuring data correctly. For the adoption of the graph to be successful, it is therefore important to approach the problems by giving priority to the relations between objects.
Last year, CIOs made multi-cloud their priority, a sign of their desire to keep control of their company’s IT strategy at a time when it began to use public cloud services. However, many of these projects spawned standalone applications unable to run on different cloud deployments.
In some cases, the possibility of seeing certain projects run on the public cloud and others on existing On-Premise infrastructures will be enough to be able to boast of “doing multi-cloud”.
For most CIOs, this evolution is a springboard towards the materialization of projects based on a hybrid cloud or multi-cloud infrastructure. Running strategically important applications on multiple cloud services independently and at scale presents its share of challenges that will finally be overcome in the course of 2020.
Along with this progress will be a better understanding of the concept of multi-cloud and its implications for data, the joint and complementary use of specific cloud services, and database architectures designed for multiple types of cloud.
5 – The effects of moving to containers are starting to be felt in data management practices
Databases tend to adapt to the design of applications a year or two behind schedule, which means that the spread of containers and microservices will start to be felt on the architecture choices of the databases that will see the day in 2020. It must be said that a growing number of microservices are now in the production phase, and that the architecture must evolve to allow their full development.
In 2020, developers will have to get closer to their operating teams around topics such as availability, resilience, and securing data produced by microservices. For this work to be successful, they will have to integrate these constraints from the start, which involves the choice of the type of database but also the data replication strategy.
In recent years, the Kubernetes project has been on everyone’s lips and sold out at various Meetups around the world. In 2020, with feedback from real deployments, however, it should lose its splendor. We can see the characteristic life cycle of these projects, but also the announcement of a future full of promise.
We can therefore expect his design errors to be the center of the conversation next year. New projects will also try to take its place, erasing the weaknesses of the original model. Some people are already voicing their skepticism about choosing Kubernetes in the first place.
In general, operations teams considering adopting k8s for large deployments will exercise increased caution when evaluating the advantages and disadvantages of this technology.
5G is still in its infancy. Launches have taken place here and there, and we are already seeing some very good marketing campaigns that praise consumers. However, it will be necessary to wait until 2020 to observe the rise of the first deployments and their adoption by companies.
For companies that want to capitalize on 5G is a big promise. These new data sources should help them to consider richer use cases. For example, extracting more data from each device and being able to query it at more regular intervals will allow logistics companies to gain more accurate insight into its performance.
Note, however, that this will lead to an explosion in the volumes of data to be saved and managed. They will therefore need to strengthen the capacities of their applications and their databases to absorb this torrent of data. The cost of the operation will have an impact on the design choices of services intended for connected objects.