The interview
Q: What is your name?
I'm Stéphane Varoqui, VP of Product at Signal18.
Q: What is your product?
Signal18 Cloud SAS was created in 2017 to support an open source product replication-manager that is embedding best practices to operate databases cluster on premise or in the cloud
My co-founders Guillaume Lefranc and me have a long and successful database industry background with the focus on serving MySQL and MariaDB customers.
Our main product is our Replication-Manager, which was born to provide service continuity for databases clusters using advance features around replication including proxying, parallel replication, GTID, semi-synchronous, group replication. Acting as an operator for provisioning a cluster of databases on premise or by instructing cloud orchestrators, it cover scopes like configuring, monitoring, backing up, scaling.
Q: Do you provide your technology to cloud operators?
This is our goal, but today we are more successful with helping clients inside their infrastructure to catchup with features normally only found only in the cloud.
Q: Are you a market leader? Who do you consider leader in your market? Does GAFAM play any role in your market?
Amazon, Google Cloud, Alibaba, Microsoft are established leaders and are driving innovation of the whole market.
GAFAM ramp up on open source database clustering for their internal usage and sometime bring innovation on top of the main database open source trunk. They later on expose databases continuity of service and scalability to the market via some cloud infrastructure and get rewards for shadowing the complexity. In such scenario, cloud users get locked in by the lack of knowledge for operating databases at scale and providers can push work at the database trunk or just trunk catch on a innovation to not stay behind
Few companies trying to tackle that market and to name our direct competitors they are Percona in US and Severalnines in Europe
Q: What are the unique selling propositions of your product? Is your solution different or better than competing solutions?
Our solution is unique as it covers a panel of clustering solutions, some are best suited for various usage like IOT, E-commerce, Banking. We abstract well established clustering practice used by the GAFAM, fully covered by the open source database trunk but we also expose emerging solutions that are cooking in the database main trunk, some are already used and pushed, for example by the DBASS Cloud provider from China and Japan with great success.
Q: Which free and open source software solutions do you use?
We use and drive MariaDB, MySQL, HaProxy, ProxySQL.
Q: What components did you develop yourself?
The orchestrator is developed in Golang.
Q: What are the strong points of your technology?
Our technology is easy to deploy and brings quick solutions when databases start to become your main growth bottleneck, as it was said in the Valley: "scale fast or die".
Q: Can you name users or clients of your solution? Preferably in the CAC40, DAX30, Fortune500, European governments?
Figaro, Gelber Group, Voltalis, TNSI, etc.
Q: Why did this client choose your solution rather the solution of the market leader or GAFAM?
GAFAM success is driven by having full control over their infrastructure and software stack. Our product is helping our clients to get hands back on their data scope on their own infrastructure. This is the opposite of the cloud where infrastructure is delegated and at some very high cost when complexity is rising or with market addiction.
Cloud databases market rely on dangerous hypotheses
1 - That cloud players will fight each other for the benefit of client cost saving;
2 - That cloud players will still innovate.
The wall will be after most IT workers behind open source moved under GAFAM by being the only source of possible monetization. From the time of patented software, the database market was in very few hands and it was sucking margins of so many industrial players, and slowed innovation until open source challenged such players.
Q: What European policies do you suggest to ensure sustainable development of your technology and its adoption?
Policy controlling data via controlling the cloud orchestrators and infrastructure software provider will lead to big competitive advantages in future.
For France, we should stop patenting US controlled cloud software like Kubernetes, and start pushing big players (CAP, Atos, Thales,,Orange, OVH, Airbus, Banks, Telco, Gov) for sponsorship and participating on European controlled cloud orchestrators like SlapOS and OpenSVC, We use flavors of such software inside our product.
I would also like to make a case for giving a specific support to the MariaDB database. A group of Europeans engineers, including Russians and Ukrainians, pushing together to bring more innovation on the cloud database market as reaction to Oracle and Microsoft dominance. We can clearly see that European big sponsor of such an initiative are missing. Investments come from US and China while European actors just act as consumers not willing to invest or to contribute a stone to the edifice. Just look at MariaDB sponsors:
- Alibaba Cloud - https://www.alibabacloud.com/ - Platinum Sponsor
- Tencent Cloud - https://cloud.tencent.com - Platinum Sponsor
- Microsoft - https://microsoft.com/ - Platinum Sponsor
- MariaDB Corporation - https://mariadb.com - Founding member, Platinum Sponsor
- ServiceNow - https://servicenow.com - Platinum Sponsor
- Intel - https://www.intel.com - Platinum Sponsor
- SIT - https://sit.org - Platinum Sponsor
- Visma - https://visma.com - Gold Sponsor
- DBS - https://dbs.com - Gold Sponsor
- IBM - https://www.ibm.com - Gold Sponsor
- Automattic - https://automattic.com - Silver Sponsor
- Percona - https://www.percona.com/ -Sponsor
- Galera Cluster - https://galeracluster.com - Sponsor
- Google - https://google.com - Sponsoring encryption, parallel replication and GTID
- Facebook -https://facebook.com - Sponsoring non-blocking API, LIMIT ROWS EXAMINED et
The open source code behind this technology become the standard, Open Source is the framework but visionary players participate in its evolution by bringing investement and teams work for granting a dominating position and get ability to influence a standard in one direction or another. It's happening now with Kubernetes where GAFAM are fully under control . It's so easy to flavor their interest, not including any other innovation in the trunk like some drivers from Huawei storage or disruptive technologies like ARM processors. IBM power processor was partly stopped by not having their own solution making them invest massively on infrastructure and getting OpenShift kubernetes orchetsrator via the redhat acquisition. Europe and European players are risking being left behind by neither contributing nor participating.
For anti-trust solutions, it would be interesting to have a policy that stated that in Europe, network communication should be priced the same from outside and inside a cloud infrastructure and a cloud provider that gives free credit when sponsoring client "addiction" should put forward the same amount of credit value to a European cloud foundation for anti-trust protection.
Q: To what extent are the interoperability projects financed by France and Europe likely to guarantee the development of European cloud technologies?
We spend most of our time to secure our income via helping our clients on database issues. We have proven database knowledge but don't spend enough time for developing our product. Financing would help dedicate time for improving the product and found research on areas we currently not good at like introducing AI based configuration and improve our front web interface.
Q: Tell us about a successful implementation. How did you implement it, why was it a success, where is this implementation today? Plans for the future? How is the relation with the client? Did the client help you get other clients?
We operate a cloud infrastructure for our client mixr.net, using OVH dedicated servers inside 3 regions, the stack is managed using OpenSVC and replication-manager orchestrators, and administrate the full LAMP stack using docker based services for a fix and predictable cost.