ethdns Contact

ethdns/Proposals

Proposals

ethdns is a non-commercial project. Its administrators, however, are always open to proposals, commercial and non-commercial alike. And when a task is not for us, we can recommend partners who will do it well.

Who you would deal with

ethdns is run by its administrators: people with twenty years of hands-on experience in building and running internet services. Over that time we have worked on projects of every size, from a single website to infrastructure that other businesses depend on, and with problems of every kind, from business logic and everyday trouble to deep technical failures at three in the morning.

We are not a hosting company and we do not have a price list. Anything we do together is agreed individually, in writing, with clear responsibilities on both sides.

What we can help with

  • Hosting for simple and complex projects, when the project fits and we agree on it.
  • Infrastructure: design, setup and care of servers, networks and the services on them.
  • Migrations between providers, servers or versions, without losing data or mail on the way.
  • Mail and DNS that work: delivery, authentication, reputation, clean zones.
  • Performance and reliability: heavy websites, slow databases, services that fall over under load.
  • Security: hardening, reviews of an existing setup, cleaning up after an incident.
  • Monitoring and backups that someone actually watches and tests.
  • Networks for offices and homes: routers, VPNs, links that stay up.
  • Rescue: a project that is stuck, a setup nobody understands any more, a previous admin who is gone.

Expertise, layer by layer

From the wire to the name, bottom to top. This is what the administrators have built, run and repaired over those years. It describes experience rather than a single setup: for each task we choose what fits, and we are comfortable learning what is not on the list.

Our emphasis is on open-source technology. Open code can be inspected, fixed and moved, and it does not lock a project to one vendor. We are not dogmatic about it, though: where a commercial product is the better tool, we use it.

L1Servers
Dell PowerEdge and iDRAC, Supermicro and IPMI, HPE ProLiant and iLO, Lenovo ThinkSystem (formerly IBM System x); out-of-band management, firmware and the full hardware lifecycle; copper and fibre cabling.
L1Disks and controllers
NVMe and SAS SSD from Samsung, Kioxia (formerly Toshiba Memory), Solidigm (formerly Intel’s SSD business) and Micron; Western Digital Ultrastar (formerly HGST) and Toshiba MG hard drives. RAID and HBA controllers: Broadcom MegaRAID and HBA 9600 (formerly LSI; tri-mode SAS, SATA and NVMe), Dell PERC, Adaptec SmartRAID (now Microchip), GRAID SupremeRAID for NVMe. Disk shelves and arrays: Dell PowerVault ME5, NetApp AFF and FAS, Pure Storage FlashArray (now Everpure), Supermicro JBOD and JBOF.
L2Switching
Switched networks from the access layer to the data centre: Ethernet, VLAN (802.1Q) and QinQ, link aggregation (LACP, 802.3ad), MLAG, STP and RSTP, LLDP, 802.1X, jumbo frames. Switches from Cisco (Catalyst, Nexus), Arista, Juniper (now part of HPE), MikroTik and Ubiquiti.
L3Network layer
IPv4 and IPv6 dual stack, addressing and subnetting, static routing, OSPF, BGP (eBGP and iBGP, multihoming, route filtering, RPKI), VRRP, policy-based routing, NAT.
Edge and security
nftables, iptables (legacy), firewalld and pf; fail2ban; WireGuard, IPsec and OpenVPN; TLS and PKI; SELinux; hardening and incident response.
CDN and DDoS protection
Where a project needs them: Cloudflare, Gcore, Akamai, Fastly, Amazon CloudFront and AWS Shield, Google Cloud CDN and Cloud Armor, Imperva (now part of Thales). Anycast, BGP Flowspec and RTBH, scrubbing, WAF, rate limiting, caching and origin shielding.
Virtualization and cloud
KVM and QEMU with libvirt, Proxmox VE, VMware vSphere and ESXi (now Broadcom), Hyper-V; OpenStack (Nova, Neutron, Cinder, Swift); AWS, Google Cloud, Azure, Oracle Cloud, Hetzner Cloud.
Storage
Ceph (RBD, CephFS, RGW), OpenZFS, LVM and mdadm, software RAID for NVMe, XFS, ext4 and Btrfs with quotas, NFS, iSCSI; S3-compatible object storage (Ceph RGW, MinIO) and cloud storage.
Operating systems
Linux: RHEL, AlmaLinux, Rocky Linux, Oracle Linux, CentOS Stream (CentOS Linux itself ended in 2024), Fedora, Debian and Ubuntu. FreeBSD (jails, pf, ZFS) and OpenBSD, and Windows Server when a project needs it. systemd, package management, kernel tuning.
Containers and automation
Docker, Podman and Compose, Kubernetes; Ansible, Terraform (HashiCorp, now IBM); Git with CI/CD (GitHub Actions, GitLab CI); shell and Python for everything in between.
Web and applications
nginx as web server, reverse proxy and cache, Apache httpd, HAProxy, Varnish; PHP-FPM (PHP 8, Laravel, OPcache), Python (Django, Flask, FastAPI with gunicorn and uvicorn), Node.js, Go; single sign-on with OAuth and passkeys.
Data
MariaDB and MySQL (replication, Galera, now part of MariaDB), PostgreSQL (replication, PostGIS, pgvector), Redis and Valkey (its open-source fork), SQLite; tuning, backups and migrations without downtime.
Mail
Own mail servers with Postfix, Exim, Dovecot, OpenDKIM and Rspamd. Hosted mail and routing: Google Workspace (formerly G Suite), Microsoft 365 (formerly Office 365), Zoho Mail, Proton Mail, Fastmail, Cloudflare Email Routing. Sending services: Amazon SES, Oracle Cloud Email Delivery, Cloudflare Email Sending, Mailgun (now part of Sinch). SPF, DKIM, DMARC and MTA-STS; deliverability and sender reputation.
Names
DNS with BIND, PowerDNS and Unbound, DNSSEC, anycast DNS; managed DNS at Cloudflare, Amazon Route 53 and Google Cloud DNS; zones, registrars, transfers and CAA.

Across every layer: monitoring and alerting (Prometheus, Grafana, Zabbix, Alertmanager), logs (journald, Loki, the ELK stack, now Elastic Stack), traffic accounting (vnStat, NetFlow, sFlow), backups (restic, rclone, Borg, snapshots) and documentation that someone else can follow.

Product and company names belong to their owners. Where a company or product has changed owners or names, we say so in brackets and use the name the market still knows. They describe our experience and do not imply a partnership, certification or endorsement.

How it works

  1. You write to proposals@ethdns.com with the basics of your project.
  2. We reply with questions, or tell you honestly that the project is not for us and recommend a partner who fits it better.
  3. We agree on scope, terms and who is responsible for what, in writing.
  4. We get going, keep you updated and avoid surprises.

Your first message

A few lines are enough to start. It helps to know:

  • who you are, and how you found us;
  • what the project is and where it runs today;
  • what you need, and what a good result looks like for you;
  • the timeline, and a budget range if you have one.

If you already use one of our services and your proposal is about it, say which one.

What we do not take

  • anything listed under What we do not allow: spam, phishing, malware, attacks, infringement;
  • link exchanges, guest posts, advertising and SEO offers. These are not answered;
  • projects that need to stay anonymous towards us: we have to know who we are dealing with.