ayedo Edge Cloud: The infrastructure in front of the infrastructure
Anyone building highly available applications invests a great deal of effort in redundant Kubernetes clusters, distributed storage and resilient databases. Yet another failure domain is quickly overlooked: the infrastructure in front of the actual workload.
After all, DNS, routing and load balancing determine whether an application is reachable at all.
With the ayedo Edge Cloud, we therefore operate these functions as an independent infrastructure layer. The key is the separation of edge and compute: the backends do not have to run in our cloud. They can run at ayedo, with another provider, across multiple clouds or on premises.
This also makes it possible to use the edge layer to bring different infrastructures together behind a common entry point.
Anycast instead of individual entry points
The technical foundation is our multi-homed edge network with multiple points of presence in the EU.
Traffic is routed via Anycast. The same IP address is announced from multiple locations, and routing determines which PoP a connection is sent through based on the available network paths.
This has an important advantage over architectures where failover is primarily implemented via DNS: if a location becomes unreachable, there is no need to first update a DNS record and then wait for TTLs to expire.
If an edge location fails or can no longer be reached via a particular path, routing can send traffic through another available PoP.
On this infrastructure, we provide Layer 4 load balancing via Anycast. TCP connections can be accepted across multiple edge locations and then forwarded to the configured backend systems.
Backends stay where they belong
Multiple backend endpoints or backend pools can be configured behind a load balancer.
The Edge Cloud continuously checks their availability using health checks. If an endpoint is no longer reachable, it can be removed from traffic distribution.
This is particularly interesting for distributed architectures.
For example, a company can operate Kubernetes clusters in different data centers and place a common edge layer in front of them. Existing on-premises infrastructure can also be integrated without first having to migrate the actual workloads to ayedo.
For us, this decoupling is an essential part of the architecture.
Edge infrastructure should not create a new dependency on the underlying compute platform.
DNS, load balancing and WAF at the edge
In addition to load balancing, we provide Anycast DNS. DNS queries can therefore also be answered across multiple regions instead of depending on a single location.
For HTTP(S) workloads, a Web Application Firewall complements the edge infrastructure. Protection mechanisms can therefore take effect before a request reaches the actual cluster or data center.
The Edge Cloud combines several functions that are often considered separately:
DNS provides name resolution. Anycast brings the connection to a reachable edge location. The load balancer distributes it across available backends. Health checks determine which endpoints actually receive traffic. The WAF can inspect HTTP(S) traffic before it reaches the backend.
The application itself remains independent of all of this.
Not just for the browser
Another important difference is at Layer 4.
The Edge Cloud is not designed exclusively as a reverse proxy for traditional HTTP applications. L4 load balancing also makes it possible to provide TCP-based services via the edge infrastructure.
Proxy Protocol v1 and v2 are also supported. This allows information about the original connection to be passed to the backend even though the load balancer sits between the client and the application.
For operations, endpoint monitoring and Prometheus metrics are available. This allows the state of the edge and backend infrastructure to be integrated into existing monitoring and observability environments.
For dedicated setups, custom IP prefixes can also be integrated via Bring Your Own IP (BYOIP). Companies can therefore use their own address ranges instead of tying external reachability to the platform operator's IP addresses.
Shared from EUR 49.95 per month
We offer the infrastructure in both shared and dedicated models.
The shared load balancer starts at EUR 49.95 per month. This includes 1 TB of traffic, up to 10,000 connections, EU-wide Anycast routing, health checks, endpoint monitoring and Prometheus metrics.
This makes the shared model suitable even for applications where operating a dedicated global or European edge infrastructure would hardly make economic sense.
Dedicated clusters are available for larger environments and requirements for dedicated infrastructure.
Sovereignty is also a question of architecture
The discussion around digital sovereignty often focuses on where data is located and where a cloud provider comes from. Both are relevant, but technically they do not tell the whole story.
What also matters is how many components of an architecture are forced to depend on one another.
If DNS, edge, compute and storage only work together with a single provider, a significant technical lock-in is created, regardless of how portable the container itself may be.
That is why we deliberately separated edge and compute.
A Kubernetes cluster can be replaced. A backend can move. Another location can be added. The edge layer does not have to be rebuilt as well.
Conversely, using our Edge Cloud does not force anyone to run their compute infrastructure at ayedo.
For us, digital sovereignty does not mean having to operate everything yourself. It means being able to make architectural decisions and change them again later.
That is exactly why we built the ayedo Edge Cloud: as an independent layer for DNS, routing, load balancing and protection in front of the actual workloads.
We are looking for infrastructure partners for the ayedo Compute Cloud
We are continuing to expand the ayedo Compute Cloud and are looking for infrastructure partners in Germany and Europe.
As part of a current market exploration, we would like to talk to IaaS providers, data center operators, MSPs and system houses that have their own or available compute infrastructure and are interested in providing it together with ayedo as a cloud offering.
The principle: You operate the data center, hardware and virtualization. ayedo takes care of the cloud layer above it – from orchestration and Kubernetes to networking and lifecycle management, billing and the end-customer portal.
We are explicitly interested not only in the large and well-known cloud locations. In particular, we want to find out which regional and European compute capacities are available and how we can integrate them into the ayedo Cloud.
To do this, we have launched a structured market exploration. We are asking for information on locations, certifications, compute, virtualization, networking, storage, provisioning and indicative commercial terms, among other things. Partial responses are also welcome.
Do you operate suitable infrastructure, or do you know a provider we should definitely speak to?
👉 Take part in the market exploration:
https://portal.ayedo.cloud/markterkundung-compute-cloud
Participation is non-binding and confidential. We look forward to exciting new locations and partners for the ayedo Compute Cloud.
The Brand Team is going on a business trip.
Around a year and a half ago, we picked up on a social-media trend at the time and turned it into a small bet with our management team:
If one of our videos reaches 3,000 likes, our Brand Team gets to spend three days in Mallorca.
We not only reached the goal, we exceeded it by a wide margin:
The video in question reached almost 7,000 likes on TikTok.
That meant the bet was won – and now it is official:
In October, our Brand Team is heading to Mallorca for the closing.
What started primarily as an entertaining social-media campaign also shows, in retrospect, how much our content has developed over the past few months.
Today, our videos regularly reach several hundred thousand people per month. In particularly strong months, our reach is even as high as around one million people.
This development shows us that authentic content, a distinctive style and the courage to try new formats can work.
Of course, Mallorca will not be only about closing and celebrating.
As the Brand Team, we will use the opportunity to produce new content, try out formats and cover the trip in our communications. The cameras will of course be coming with us.
After a year and a half, a small social-media bet has therefore become not only a trip, but also a nice opportunity to look back at the development of our Brand Team and our social-media channels.
So for three days in October, it will be:
Office for Mallorca, everyday routine for closing – and, of course, plenty of new content.