This blog post will drill down on why Metal Instances play a crucial role in the enrichment of our compute portfolio, outline the key benefits that this new service brings to the market, and explain how organizations use such a services.

Metal Instances contribute to our IaaS and compute portfolio enrichment, offering more choices to customers

If you are already using our Public Cloud, you are likely aware that you can choose from a broad catalog of instances ranging from General Purpose, CPU-optimized, RAM-optimized, GPU, or I/O intensive instances to get the resources that you need on-demand and consume on a pay-as-you-go model. You might even be automating the deployment of these resources leveraging Infrastructure as Code tools such as Terraform or Ansible which are natively supported as our Public Cloud is built upon standard Openstack.

When looking at the bigger picture, the choice of instances is related to application workloads; not all apps are created equal and some workloads might require dedicated hardware resources for their execution, whether it’s for control, isolation, or performance. This is where Metal Instances come in. You can get a single tenant node with dedicated hardware resources and combine it with the on-demand availability and quick deployment times based on automation.

For a quick view of how Metal Instances fit into our overall compute portfolio, refer to the table below.

Metal Instances bring together the best of Bare Metal and Public Cloud

Now, let’s review in more detail what the benefits are for you.

• Single-tenant node, as a Service

There isn’t a virtualization layer installed on Metal Instances. You get root access, meaning that you have access to the highest privilege mode and can install any low-level software. All hardware resources available on the Metal Instances are dedicated to your usage and you control resource allocation.

• Integrated into the entire Public Cloud ecosystem

A simple way to look at Metal Instances is to view them as an additional building block for your app. As with any Public Cloud service, Metal Instances are consumed within a Public Cloud project where you can consume storage, network, databases, orchestration, and even off-the-shelf AI services.

• Available and scalable on demand

Getting what you need, when you need it is one of the key promises of our Public Cloud. Metal Instances can be deployed in a few minutes within your project and can be horizontally scaled up and down to match your workloads over time.

• Billed hourly and/or monthly

The benefits of pay-as-you-go have already been largely advertised, but are worth covering again in the context of Metal Instances. You pay for the hours used if you turn to hourly billing, which offers great flexibility for temporary workloads. On the other hand, if your workloads are permanent, monthly billing is another option.

• Natively compatible with Infra-as-code tools for full deployment automation

Infrastructure as code provides a great productivity advantage, as it saves your operations team from performing repetitive tasks when it comes to setup. Popular tools such as Terraform or Ansible are natively supported which makes things easier for standardized infrastructure deployment.

Key takeaways from our beta customers

About one hundred customers have been beta and alpha testing Metal Instances over the past months and feedback about their use cases provides insight into how you might use this new service.

Automation and industrialization

Several beta customers had the need to be able to install KVM and have access to the underlying tooling while being able to industrialize deployment as a best practice. If you’re already using an Infra as Code tool, Metal Instances can be a great time saver and there are a number of use cases around development or CI/CD which are worth considering

“We need to be able to add capacity at peak times automatically through an API. This does not work with regular dedicated servers that need 1-2 days of setup. We need new capacity in minutes“.

Specific software licensing

For licensing reasons, some applications require specific low-level software. We’ve seen this case with some beta customers that configure a Red Hat virtualized environment to run an Openshift cluster. The requirement is twofold—a certified hypervisor is required to run Openshift, but to standardize deployment, automation is needed too. Another case was from a customer moving to the cloud and concerned the installation of Windows Desktop in a manner to comply with the licensing for the application running atop while being able to reuse its automated workflow in the cloud.

Video related use case

With all hardware resources being dedicated to the workload, some beta customers wanted to implement video-related use cases such as conferencing or rendering.

“Metal Instances are a great solution for our heavy workloads because they give large amounts of resources for short amounts of time. We need to be able to increase our capacity during the day and vice versa at night.“

Typically, the fact that Metal Instances are available on demand, billed hourly, and can be automated, makes them a perfect solution in that light.

Wrapping up

The launch of Metal Instances is a key milestone and our teams that made this release possible deserve huge congrats! To get the full details about the products, go to our webpage: https://www.ovhcloud.com/en-ie/public-cloud/metal-instances

The table below provides the specs for Metal Instances. Hourly prices start from 0.5€ and monthly price starts at 170€ and the detailed price list can be found here: https://www.ovhcloud.com/fr/public-cloud/prices/#419

Currently, Metal Instances are available in our Beauharnois, Strasbourg, Gravelines, Limburg, Erith, and Warsaw datacenters. Stay up to date with location availability here.

For new customers to our Public Cloud, you can take advantage of our free trial and test this new service along with our full catalog. We are always looking for feedback, feel free to reach out on our Discord channels. You may also want to check out our publicly available roadmap on Github and propose new features and interact with our Product Managers.

Our customers have been able to build, deploy and run their applications in the cloud by leveraging our Public Cloud, a coherent and integrated ecosystem of infrastructure and managed services. As our customers’ cloud applications evolve, the need for ready-to-use, managed services has become more prevalent, as it enables them to focus on adding more capabilities to their apps. While customers have outsourced to cloud providers the responsibility of managing foundational infrastructure services, they are looking for services that make it easier for them to concentrate on their business logic and network services are no exception to this trend. In this post, we’ll highlight the new services we are building to meet the needs of these types of customers.

Network managed services

Though the network is thought of as an infrastructure component, it is interesting to differentiate the raw capability of providing network connectivity from the network-related services like routing, load-balancing, and traffic management. In the past, it was possible to add these capabilities on your own using compute instances and deploying software on top of the components that perform the aforementioned network-related services. The main benefit of the services discussed in this post is that customers no longer need to implement and maintain such services, as we now provide them.

Load Balancer

While our Kubernetes customers have been able to use our Load Balancer, integrated with our Managed Kubernetes, we now provide a network and application-level load balancer compatible with our Public Cloud Openstack deployment. This load balancer is built upon the OpenStack Foundation’s Octavia project and can be managed through Openstack tools (API, CLI, or Horizon). This is a great fit for users that are familiar with the OpenStack environment and are looking for an infra-as-a-code approach. Load Balancer is available in a variety of sizes to accommodate different usage and performance levels and provides typical cloud features: dynamic public or private environments, high availability, and an interface to cloud-native platforms. Check out all the details on our product page: https://www.ovhcloud.com/en-ie/public-cloud/load-balancer/.

Gateway

Enabling to connect networks, Gateway brings more flexibility to manage resource access through the network. A typical Gateway use case is to split your architecture on multiple networks and configure routes on demand. Gateway can also be an internet gateway, with instances using private ports to remain isolated from the public. In addition to classic private network management tools, Gateway brings more granular management to your networks. More details here: https://www.ovhcloud.com/en-ie/public-cloud/gateway/

Floating IPs

Floating IPs are public IPs that can move from one instance to another and are handled by Gateway to forward traffic to the assigned public IP. There are several situations where floating IPs are useful:

# You can define a floating IP for your application in advance, without having to create an instance to get a public IP. This can help you manage DNS records.
# You can link your floating IP to instances reachable through your virtual router, allowing you to enable internet ingress traffic on-demand.
# You can switch the link from one instance to another at any time, which can help you manage outage situations, or simply schedule maintenance actions (running rolling upgrades on multiple nodes, for example). Find out more about the product here: https://www.ovhcloud.com/en-ie/public-cloud/floating-ip/.

What’s in for me?

Here, we’ll take some concrete examples so that you can better understand what each service can do for you. Floating IP, Gateway, and Load Balancer can be deployed together to create the appropriate rules for network accessibility while providing the required security. The following scenarios highlight the key benefits of each service used in combination.

Floating IP scenarios

Gateway scenarios

Load Balancer scenario

Next steps

A big congrats to our teams that made the release of our Load Balancer, Gateway, and Floating IPs services available to our customers! At launch our network services, are available in the following Public Cloud locations: GRA9, GRA5, GRA7, SBG5, SBG7, UK1, DE1, WAW1, BHS5, SYD1, SGP1. After the launch, keep track as additional locations are rolled out here.

For new customers to our Public Cloud, you can take advantage of a free trial and test these new services along with our full catalog. We are always looking for feedback, feel free to reach out on our Discord channels. Last but not least, check out our publicly available roadmap on Github and propose new features and interact with our Product Managers.

Hence, the start of this Public Cloud changelog. We’ll communicate regularly about the latest features to update our customers. For the sake of consistency, we will organize the section of this post, in the same way our roadmap on Github is organized. We hope you’ll enjoy it and find it useful. Don’t hesitate to share your feedback!

Infrastructure: Compute, Storage and Network

• Our entire OpenStack infrastructure has been updated from Newton to Stein version. Beyond versioning, the support of OpenStack Stein brings a host of improvements that will enable us to deliver new capabilities. For example, on the networking side, you can now leverage Security groups to manage network access rules for instances. More is coming soon on the network, so stay tuned on this.
• On the Storage front, in case you have missed our announcement, our Standard Object Storage – S3 API is now available in Gravelines and will be rolled out our other datacenters next. Providing complete S3 API coverage, you get the best price on the market, for storing any static file with unlimited scalability.
• This release complements our Object Storage portfolio, where High Performance Object Storage – S3 API is available in our Beauharnois datacenter. For reference, you can check the available list of services in our datacenters here.
• Last but not least, Object Storage – S3 API now supports server side encryption and S3 object lock, thus bringing more advanced security features.

Databases

We now offer more than 12 Database Management Systems, available as a managed service and billed upon usage within our Public Cloud. The key highlights from the latest months include:

• For all DBMS we do provide Terraform integration, easier connections with code examples and Docker and backup and maintenance timeslot adjustment. Check the video below for more details.

• For all DBMS, we have improved order ergonomy, so you can more intuitively and quickly select the sizing and features of your managed database cluster.
• We have added new flavors to all our DBMS catalog, now starting from 4GB RAM per node
• For MongoDB DBMS, we provide an easier way to connect with DNS Seedlist features

And as a video is worth 1000 words so have a look at the control panel to see those new features in action.

Containers & Orchestration

A growing numbers of our customers are now adopting a cloud native approach for their applications. Most of the new features we have developed over the past months are related to our Managed Kubernetes Service. Below are details about the purpose of each new feature.

Security

• Clusters spawned in vRack (OVHcloud cross-region private network) can now specify a custom Gateway IP (or use the one declared through OpenStack private network DHCP) to route all worker nodes’ traffic.

Automation & Scalability

• CoreDNS configuration can now be supercharged with custom entries that will be preserved during a cluster patch or upgrade.
• Following user feedback, we made the DesiredNodes property of Node Pools optional.
• We improved our Terraform Provider, assuring that Kubernetes resources cover our entire feature set.
• You can now choose the value of AlwaysPullImages

Kubernetes Lifecycle

• We released Kubernetes 1.24 for both new and upgrade to existing clusters.

Other key updates

• OVHcloud obtained KCSP (Kubernetes Certified Service Provider), meaning that with multiple CNCF certified experts, our Professional Services team can help enterprise customers in their cloud native journey with OVHcloud.
• We published new tutorials on policy management with Kyverno  and  Labels & Taint on Node Pools and updated more than 20 Kubernetes documentations pages.
• Next month, we will make a small breaking change to the NodePool CRD, giving you more precision when a NodePool is being resized.

AI

Our AI development team keeps on rocking and delivered many great features, including:

• What you used to know as AI Apps, has been renamed to AI Deploy and we released an Alpha version release.
• We are now offering on AI Notebooks a framework that can run quantum algorithms
• New python, open source libraries for machine learning, computer vision and image processing are proposed by default in AI Notebooks
• On AI Tools, we have improved rights management with “Read Write Delete” granularity
• Stand-alone volumes are available on AI Tools, so the experience is improved with framework that are using a lot of temporary files.
• We have released many new tutorials to make it easier to use AI Tools. Check them out here.

Too Long Don’t Read ? View those updates in video demo.

Let’ say, for an AI-based application, you want to use GPU to make some experiment, design a prototype or start your business: You get the GPU resources as you need them, and pay only for what you have used. The technology you need is readily accessible, thus shortening your time-to-market, while you don’t have any upfront investment.

In this light, this post will review the different ways at your disposal, to consume GPU-powered services in our Public Cloud, and highlight how to leverage them in the context of building data-intensive, AI-based application.

GPU(s) are a great fit for AI workloads

As its name states it, Graphical Processing Unit’s straightforward goal is to efficiently manipulate computer graphics and image processing. GPUs are equipped with many cores, provide high throughput, thus have much better ability than CPUs to efficiently perform parallel computing tasks and process thousands of operations at once. These capabilities can apply to many use cases ranging from gaming to high performance computing, but more notably to AI.

In short, thanks to its ability to process multiple computations simultaneously, GPUs have greatly answered the needs for Data scientists: for instance when training a Deep Learning model in conjunction with large data sets, doing it with a GPU can be completed in a few hours, while it could last for weeks with a CPU.  

GPU-based instances: Our T2 instances

The first way to use GPU in our Public Cloud is our GPU instances. What you get is a Virtual Machine, that you can provision on-demand, manage through APIs, and whose CPU resources (vCore, RAM) are dedicated to you. But most of all, you get GPU resources, from one to four Nvidia Tesla cards, delivered directly to the instance via PCI Passthrough: As there is no virtualization layer, your AI-application workloads can directly access the GPU, get the whole available computing power and benefit from outstanding performance.

For running GPU instances, you can now provision our T2 instances (for Tesla 2nd generation) from your control panel or through APIs. In terms of specifications, those T2 instances are built with the Nvidia Tesla v100s GPU card, which provide 32GB for RAM per card (vs 16GB of RAM for T1) and have more CPUs core than T1, as shown in the table below.

And, of course T2 instances do follow our SMART approach:

• Simple pricing to understand
• Multi-local: available in Gravelines, Beauharnois and Limburg datacenters
• Accessible: well… check the price of p3.2xlarge to make your own idea
• Reversible: just take your data back and delete your instance
• Transparent: no ingress, no egress so it’s highly predictable

A few tips

1. You can easily deploy GPU-accelerated containers on top of our T2 instances, with the Nvidia GPU Cloud (NGC) catalog.
2. Make the most of the pay-as-you-go. Just delete your T2 instance when don’t need it anymore, as you’re charged per hour of usage.
3. A way to look at our T2 instances is to consider them as an alternative and complementary to your GPU server without the upfront cost.

But wait, how about using GPU power with ease, with no drivers to install, no linux to update, with automation and AI framework and notebooks ?

Glad you asked !

AI Training: Train your models with GPU(s)

While you get the freedom, to install whichever distribution, AI framework and tools on top of our GPU instances, doing so requires system administration and ops skills that data scientists may not have, but most of all this is not where they do provide the greatest value.

With our objective to provide ready to use solutions for building AI-applications, we are proposing AI Training which enables you to train your models in just a few clicks, or in the command line interface. AI Training is a managed service that orchestrates the GPU resources (our T2) for training your model, hence freeing your data scientists from infrastructure and maintenance operations.

The beauty of this service is that you pay only for the execution of the training job you have specified. Check the video demo to see how AI Training fit into setting up an AI pipeline.

Steps to launch a job with AI training:

1. Select your datacenter region, as of today Gravelines and Beauharnois
2. Enter your Docker image, basically the model you want to train. The image can come from public or private registry
3. Add data set against which your job will be trained. You need to create a Object Storage container to add a data set for your job.
4. Enter the Docker command
5. Configure your job (you can add up to 4 GPU for your job, but you can as well use CPU). The price per hour is displayed, but billing granularity is per minute
6. Launch the job through the control panel or through command

Once the job is finished, save your trained model and delete the job. You are charged for the number of minutes you have used AI Training to perform your job.

AI Notebooks: Code live on top of GPU(s)

AI Notebooks is the latest of our AI services in our Public Cloud. It enables data scientists and developers to deploy, within 15 minutes, a ready to use notebook providing the required GPU resources (it supports CPU as well) for live coding and for creating their own models. AI Notebooks proposes Jupyter or VS Code notebook, the two most popular open-source live code editors, and supports natively the widely-used frameworks such as TensorFlow, PyTorch, Scikit-learn, MXNet and Hugging Face.

In short, AI Notebooks simplifies the life of your data scientists and improves their productivity, as they can focus on developing models with the tools they are already familiar with. Similarly to the Public Cloud services we reviewed in this post, AI Notebooks is a pay as you go solution. Based on how many GPU resources your AI Notebooks is running on, you will be charged for the number of minutes you have used your notebook.

Wrapping up

This review was written with the goal of providing a high-level understanding of the GPU-based services we do offer on our Public Cloud. While GPU instances, AI Training and AI Notebooks do have in common the fact that they are built upon our T2 instances, leveraging Nvidia Tesla v100s, the way we have designed and packaged those services makes them very different in terms of use case, target users and way of consuming the GPU resources. You can as a takeaway from this post check the summary table below:

At OVHcloud, we are proposing Object Storage in our Public Cloud universe and we have released, last year, compatibility with the S3 API, the market standard. Simply put, organizations and developers in particular, can reuse the tools (SDK, libraries or plug-in…) they are already familiar with, to migrate their data from any S3 bucket to OVHcloud Object Storage, and benefits from the cost-effectiveness, reversibility and privacy by design of our service.

This blog post details how Systran, a pioneer and global leader in translation solutions, has migrated from AWS S3 to OVHcloud Object Storage, and will cover the drivers behind that migration and the benefits. Huge thanks to Jean-Sellenart (CEO) and John-Paul Barraza (CIO), at Systran for helping us putting together this post.

Systran technology hosted at OVHcloud

Why does storage play a key role in Systran’s business? What are the reasons for migrating to OVHcloud Object Storage?

For Systran, storage is considered to be very important, as it hosts business-critical data such as translation models and models-related data. This is vital for the business in terms of Intellectual Property, as well as affecting the data transparency and privacy promise Systran makes to its trainers. As a result, trainers expect robust protection for the models they will list on Systran’s marketplace, as well as sole ownership over their data and reversibility; their biggest concern is that the models could be used for a different purpose than what they were designed for, which would lead to loss of earnings for them.

As a software solutions provider, Systran monitors its production costs closely, as they affect its margin. Systran was initially using AWS S3 as its service for storing models and related data, but they realised that the billing was unpredictable, too complex to understand and, most of all, expensive. As Systran’s marketplace gained more traction in the market, their storage volume needs were increasing. At the same time, it wanted to have more control over costs, while increasing its protection of its business-critical assets. As Systran was already an OVHcloud customer, it naturally chose to migrate its data to OVHcloud Object Storage to address those issues.

What advantages does Systran gain from migrating to OVHcloud Object Storage ?

The first resulting benefits are cost reduction and predictability. In fact, payback was achieved only two months after the migration. The total cost of ownership (TCO) with AWS S3 would have been around $13k for three months, while with OVHcloud Object Storage the TCO, including migration cost, was $11k for the equivalent time period. The benefits of migrating to OVHcloud from a financial perspective is becoming even more apparent by the month, as the bill with AWS S3 was around $4k per month, which is more than 250% higher than what Systran is paying for the same volume of data with OVHcloud Object Storage. Furthermore, as its storage volume is increasing by an average of 10TB every month, the pricing model of OVHcloud Object Storage is simple, predictable and transparent, which is critical for Systran in terms of better controlling its costs.

From a more functional standpoint, Systran was able to streamline its storage needs, as the migration was an opportunity to delete redundant translation models, as well as provide more protection and resilience to the data stored in Object Storage. While utilising the triple replication provided by OVHcloud Object Storage, Systran also decided to archive its translations in the OVHcloud Cloud Archive to increase its business resilience. Therefore, with OVHcloud, Systran can deliver on its promise of data privacy (thanks to OVHcloud’s GDPR compliance), reversibility and resilience for its ecosystem of partners, while protecting Intellectual Property.

Finally, Systran’s migration to OVHcloud Object Storage was facilitated by the provision of standard APIs. OVHcloud Object storage offers both S3 and Openstack Swift APIs which are widely adopted standards for development, meaning that developers can use tools that they are already familiar with. Another technical advantage of OVHcloud Object Storage is the network performance, which Systran has noted to be faster than AWS S3. This means that training translation models are faster for Systran trainers, resulting in further cost reductions.

Who is Systran?

Systran is a pioneer and leader in translation solutions, serving global organisations with its Machine Learning technology. With a key mission to provide tailored translation solutions to vertical industries, Systran has launched its marketplace: a community-based platform providing specialised translation models in various fields (such as IT, legal, and manufacturing), with full data transparency and privacy. This platform lets developers of specialised translation models (called “trainers” in Systran’s terminology) list their own model(s) on Systran’s marketplace, and end customers can buy those models for use on their own business applications or website. Systran collects the payment from the end customers, then pays trainers for the purchased models.

Towards the end of 2016, Systran was once again breaking new ground, launching the first professional neural translation engine. This program makes use of the latest advancements in deep learning technology to improve the quality of instant translation. If you want to know more about them, see this use case.