Thread by @BastienOvh

Over the last months, we improved our @OVHcloud Public Cloud ecosystem with our vision + your requests and feedbacks.
Fasten your seatbelt, here is a recap ! 🔽

Reminder: Public Cloud = bricks like compute, storage, network, PaaS, services,... sharing some concepts like pay-what-you-consume-only, usually mgmt by project, API everywhere and sharing physical hosts between multiple customers. But your VMs (called instances) are for you only

Let's start with compute. We now provide IOPS instances with high-perf SSD #NVMe disks.
#ELK, #Hadoop, #NoSQL ? Perfect uses-cases for them :) Trust me they "dechire grave" as we say here.
FIO benchmarks vs competitors = specifc thread soon :)
ovh.ie/public-cloud/i…

Still for compute, new lab : Baremetal instances.
Powered by @Openstack #Ironic, it's instances without hypervisor neither virtu layer, but with flexibility of cloud. Same tools, same automation (and yes some limitations too).
Uses cases / details :labs.ovh.com/bare-metal-on-…

For storage, we are now compatible with AWS S3 API. We provide the s3api middleware, allowing you to use the official S3 client, then push in @OVHcloud Object storage (powered by @Openstack #Swift).
Tech doc to use it : docs.ovh.com/gb/en/public-c…

Let's switch to containers and orchestration. We now provide the latest version of #Kubernetes, v1.16.
And lifecycle policy is officially defined here : docs.ovh.com/ie/en/kubernet…

Also, for containers, the new lab Private Registry allows you to manage your docker images or helm charts.
Of course you can use it for all our cloud products, not only for #kubernetes
Everything you need to know is here : labs.ovh.com/private-regist…

On the Data field, you can now deploy our secured @hadoop clusters in more regions (FR, UK, DE, PL). You need 2 minutes to launch it in control panel, then approx. 1h for the complete setup (Full kerberized cluster).
Tech doc : docs.ovh.com/gb/en/analytic…

A (free) new lab appeared to provide message queuing service, it's called ioStream. For now you get an @apache_pulsar endpoint, and @apachekafka is planned. All you need to start is here : labs.ovh.com/iostream

In the Santa Claus sack bag : ML Serving Engine, a lab to easily put in production your #MachineLearning models (#Tensorflow, #PMML, #ONNX), with API endpoint, versioning, scaling, monitoring and more. Details here : labs.ovh.com/ml-serving-eng…

Another lab for data processing, will pop out soon, to launch @ApacheSpark jobs effortless. Stay tuned..

Let's move on to the features. You can now easily automate backups via control panel or API.
Select an instance, amount of backups to retain, and go. Powered by @Openstack #Mistral, you can of course create your own worfklows.
Official doc : docs.openstack.org/mistral/latest/

On security side, new feature to assign multiples roles per Openstack user. Demo done with control panel, but of course it works from API :)

... That's all folks 😎 *drop the mic* I did not speak about overall quality / QoS, but it was improved considerably. if you feel it as a customer send your hugs to @damienatovh , @pierrelibeau, @LabedzAtOVH and their teams :)

Big hug to all the people from @OVHcloud accross the world who worked on these projects & improvements !
And as you can see, using open source is the right way to do it. Many thanks to all the OSS communities, especially @TheASF and @Openstack for this time. GG !