The business of Colocation includes the housing service for IT equipment in a Data Center within this scope:
- Busy space with racks containing IT equipment.
- The electrical energy consumed by this equipment, supplied with high level of availability.
- The assurance of the right levels of temperature and humidity in the room where the IT equipment is housed for its correct functioning.
- Data connection between racks and IT rooms in the Data Center.
- Connection to carriers’ networks available in the Data Center.
- Simple operations in the equipment under the clients’ instruction.
The beginning of colocation services and their development in recent years
Thus, we can define this Colocation services as the basic level of service in a Data Center Infrastructure which is consumed by every client.
Traditionally, Colocation service providers have offered their service to a wide spectrum of clients:
- SMEs: housing of 1 or more racks with a power supply between 1,5-3 KW/rack.
- Big enterprises: housing of 10-50 racks with a power supply between 3-6 KW/rack.
- Integrators/Outsourcers: their consumption of services equals the big enterprises.
All these clients demand the Colocation services aiming to house their IT immediately in a Data Center, asking for a price based on the following: physical space availability in the geographic area required by the client; type and level of redundancy in the Data Center (TIER certifications); number of racks to be housed; power supply assigned per rack; and contract length.
The Colocation services standards have been, traditionally, set by the Colocation provider, according to its catalogue.
Nevertheless, the rapid increase of public cloud services, the size of the platforms sustaining these services and the geographic expansion of public cloud nodes deployment, have meant the burst of a new client segment in the industry of Colocation. These new clients are the global cloud service providers, also known as hyperscalers: Amazon Web Services, Microsoft Azure, Google Cloud, Ali Cloud, Oracle, Huawei… a few technological giants that vastly dominate the public cloud.
The first public cloud nodes were deployed in the United States, giving service to developers and Internet companies capable of connecting to the cloud from any geography. As the number of clients, its service consumption and the service spectrum grew, these nodes’ dimensions increased towards enormous levels compared to the traditional Colocation clients.
The natural evolution of hyperscalars within the industry
This need of further space and power supply to deploy new cloud nodes meant hyperscalers had to build their own Data Centers when there was not enough offer of Colocation services to satisfy their needs. Consequently, hyperscalers have expert teams in building and operation of Data Centers within their organizations.
Thus, public cloud services and the deployment of nodes have entered a new phase of expansion due to the following reasons:
- The need to deploy regional and/or domestic nodes to comply with the Data Protection regulations in each country they operate.
- The need of multiplying the nodes to bring the platforms near to the clients which require low latency levels to use their applications.
- The strategy adopted by hyperscalers of deploying 3 availability zones (AZ) in one city, to assure resilience and redundancy.
- The race to fulfill the 3 previous conditions before their competition.
Precisely, this urgency for making new deployments has made hyperscalers to look for Data Centers either existing or in construction phase.
To sum up, we find ourselves in a sweet moment for Data Center service providers in terms of demand. This demand, though, forms a new typology, with very demanding clients and a new relationship between provider and customer in comparison with traditional Colocation. Next, we summarize these differences:
| Traditional Colocation | Colocation for hyperscalers |
| Service from the providers catalogue | Service quality defined by hyperscaler |
| The provider offers its free space | Hyperscalers prefer dedicated space under new construction |
| The client purchases the services they need in each moment in a rank of KW | Hyperscalers purchase in MW rank, and negotiates bookings of future MW |
| The decision is made on quality-price basis | The decision is made based on space and power expansion capacity |
| Operative procedures are set by the service provider | Operative procedures set by service provider, but agreed with client |
| Uptime Institute’s TIER certificate as quality standard | Hyperscalers evaluate Data Centers under their own criteria |
| 1-3-year contracts | 10+-year contracts |
| Power density: 1-4 KW/rack | Power density: 4-12 KW/rack |
In conclusion, we face a huge new demand of Colocation services brought in by hyperscalers, which will be present during the next 3-4 years. To be able to satisfy this demand requires to develop a totally new business relation from the Data Center service providers and the will to implement large investments for new build outs with a power density unknown to us. This new environment can be the key to accelerate the consolidation process the Data Center industry is immersed in.
Eduardo Gómez-Leal, CCO & Strategy Development – Nabiax