ungleich
Infrastructure
Availability

Infrastructure availability

This is version 1.0 of the ungleich infrastructure availability definition.

Depending on your project or business case you have different availbility requirements. At ungleich we accomodate the range from "can be offline for a while" up to "is mission critical". We orientate ourselves on the It standard availability definitions "availability per year". The following table gives an overview of typical availability rates:

| %      | Accepted downtime / year |
|--------+--------------------------|
| 98     | 175.2h or 7.3 days       |
| 99     | 87.6h or 3.65 days       |
| 99.9   | 8.76h                    |
| 99.99  | 0.876h or 52.55 minutes  |
| 99.999 | 5.25 minutes             |

General availability

If not otherwise specified, the target availability for every service is 99.9%. This availability is not guaranteed by default. However we offer guaranteed availability for a range of our products.

Guaranteed availability

We can guarantee availability of individual services depending on your needs. As every customer need and product is individual, you can ask for an individual offer for your services.

High Availability Methods

To improve the default availability, we offer a variety of add-ons. Depending on the booked service, different methods are applicable. In this section we describe the general methods we offer.

Redundant storage

By default all data is replicated 3 times accross our decentralised storage system. This protects against disk failures like a traditional RAID system does. Note that this does not protect against unintentional deletion, for this you require a backup.

Offsite backup

To allow disaster recovery, we offer offsite backups for most of our services. The data is stored in a physical different location so that even physical damage to a data center will allow you to restore your data.

Dedicated Internet uplinks and IP blocks

To protect you against DDoS and impacts on our general infrastructure, we offer dedicated Internet uplinks that are reserved for your service. This includes a /48 IPv6 address block by default for BGP announcement. A /24 IPv4 address block for indidivual BGP announcement can be added at an extra service cost.

Redundant services / High availability (HA)

We realise high availability (HA) using redundant services. These services can be placed in the same geographic or different geographic data centers. Depending on the service architecture we offer Cold-Standby, Hot-Standby or active-active services.

With different geographic locations separate data storage is included by default. For services in the same location, an optional separate data storage is also available.

Load Balancing

For services that allow multiple backends, we offer load balancing services to distribute the traffic. This can be combined with dedicated uplinks per location, indivual IP address blocks and distributed application deployments in multiple locations. You can also easily realise blue green deployments or Canary releases.

Related pages