Document toolboxDocument toolbox

(sv) Data Center Information

General Information

To provide uninterrupted service, HDC.cloud is operated on professional grade hardware and software platform. Each customer can benefit from this platform for a fraction of a price when compared to in house solutions.

Our servers are located within high-security premises with both electronic and physical protective measures. Our collocation center provides fire and power outage protection and is operated by one of the leading operators in Scandinavia.

Every part of equipment is fully redundant, that means that there is no single point of failure. That includes fast internet connection, firewalls, servers, hard drives as well as cooling, fire protection and uninterrupted power supply.

Should this platform fail, due to a natural disaster or human error, there is a secondary backup facility. All data, at all times, is stored only within the European Union and accessed only by selected, authorized employees.

Non-disclosure agreements bind all Provider employees. Offsite Backup is transferred over a secured VPN connection and stored on an encrypted storage system.

Data Center Security

Data center is designed to withstand intrusion attempts, sabotage, fire etc. The facilities are CCTV-monitored and have alarm systems with perimeter and volumetric alarms. The facilities are monitored
around-the-clock by both IP-Only’s NOC and a security company.

Data Center  Physical environment

The facilities’ cooling and climate systems control the temperature and humidity in the data centers server rooms. Power supply for the server rooms is provided through UPS and backup power systems that ensure
continuous operation even during power outages in the electricity grid. The data centers are equipped with fire detection systems and have automatic fire extinguishing systems that prevent fires right at the outset.
The fire department is automatically alerted in the event of an incident.

Data Center Communication

The data centers are connected to IP-Only’s Nordic fiber network as well as other national and local fiber networks. The connections are redundant and run along separate physical paths into the data centers, i.e.
diversity. Communication path diversity is also provided within each facility, ensuring that redundancy can be achieved from IP-Only’s backbone to the customer’s server racks. 

Data Center Access

Only authorized and pre-registered people have access to the data-center. Card-key that requires a code is required, as well as a camera for ID-checks. 



General schematic for network and user connectivity:



Data Center Resiliency

Primary data center implements various solutions to minimize the risk of a downtime. Nevertheless, all of this solutions are located within one location. 

To maintain operations in case the primary location cannot be used, we have implemented an independent disaster recovery data center. (Read more about (sv) Disaster Recovery)

Application Server

Each customer is operated on a dedicated web server that runs within a virtual environment. Although called virtual, there is always a hardware host system underneath running on a physical server.

Should the server fail, each customer has a "clone" of application server residing an another server, ready to take over. Customer data is not stored on an application server so no data should be lost in that kind of situation. 

Possible consequences are that users will have to re-login and background processes (like data import) will have to be restarted.

Database

Database servers hosting customer database also run in a cluster that is being replicated. In case a host server with a customer database goes down, another physical server has a synchronized replica of data and can take over. 

User File Storage

User uploaded files are stored on a distributed file storage system. The system spans across multiple physical servers and hard drives, meaning it provides resiliency for hard drive or storage server failure. 

The file storage system provides various advanced features such as auto re-balancing and auto healing capabilities in case of hardware failures.

The system uses high performance infrastructure to provide synchronous data replication for the primary data center and asynchronous data replication for disaster recovery site.

Network 

Network equipment provides N+1 redundancy starting from a connection to a data center network provider (ISP), through firewalls and switches. As a result, each server can utilize at least two different connectivity routes. 



Energy Efficiency 

It is estimated that about 40 percent decrease in the amount of energy used by data centers could be achieved globally when moving all IT operations into the cloud.

Our data center is using efficient solutions that help to reduce the carbon footprint even further.

Facts about data center*:

  1. Environmentally-friendly power and cooling techniques

  2. Cold room with separation of warm air via chimneys and plenum

  3. Green electricity supply: 100 % hydroelectric power, GoO

  4. PUE factor 1.3 (annual average) 



What is PUE and how it is calculated:

PUE stands for Power usage effectiveness. It is calculated based on this simple formula:

PUE = Total energy entering the data-center / Energy used by IT equipment inside the data-center

DCiE is a metric used to evaluate the power or energy efficiency of a data-center. DCiE represents the ratio of total amount of energy consumed by all IT equipment and resources to the entire energy consumption of a data-center.

In case of our data-center, the DCiE metric is at 77% 



Fact about server equipment**:

  1. Our servers meet "ENERGY STAR Enterprise Server 2.1" requirements

  2. Servers are equipped with "active power management" system

  3. Energy Smart power supplies have intelligent features, such as the ability to dynamically optimize efficiency while maintaining availability and redundancy

  4. The system aims to maximize performance per watt through a combination of energy efficient technologies, optimized thermal designs and intelligent fan control algorithms.

  5.  The system fan control algorithms use an extensive array of sensors that automatically monitor power and thermal activity to minimize fan speeds based on system cooling requirements, reducing the power required for cooling.

    1. Variable speed fan control based on power or thermal readings

    2. Power capping

    3. Low power memory states

    4. Low power I/O states

    5. Processor or core reduced power states

    6. Dynamic voltage and frequency scaling of processor(s)



*) Source: DELL, ENERGY STAR

**) Source: IP-Only AB

Network Details

HDC.cloud has a guaranteed network connection performance to all of our customers. Main data-center network connections include redundant (multi link) fiber connections.

Base network parameters / performance

MTU

(Bytes)

Latency within Stockholm City

Latency within Sweden

Latency Within Nordics

Latency  Within Europe

Latency  USA

Dropped packets

MTU

(Bytes)

Latency within Stockholm City

Latency within Sweden

Latency Within Nordics

Latency  Within Europe

Latency  USA

Dropped packets

100

< 10 ms

< 30  ms

<40 ms

< 75 ms

< 110 ms

< 0,1 %

*) Note: Those are not guaranteed server API response times, only network latency.

Network speed is rated at a fixed speed of 100Mbps with burst capability up to 1000Mbps.

Infrastructure Details

Fire prevention

  • Server Rooms: Smoke gas detection with sampling and extinguishing system with inert gas/ Novec

  • Equipment rooms: Smoke gas detection with sampling

  • Public areas: Sprinkler system

  • Portable fire extinguishers in all areas

  • Sprinkler system in backup power rooms

  • Fire alarm connected to emergency services

Cooling

  • Redundancy: N+1

  • Optimal environmental control is maintained with precision cooling systems

  • Regulated temperature 22°C (+/- 5°C)

  • Regulated humidity 50 % (+/- 10 %)

Power Supply

Equipment in powered by A+B power systems. This implies two independent power channels to secure uninterrupted power supply that is maintained by battery and diesel powered generators.

  • Switchgear installed in accordance with SS EN 60439-1/-3

  • Incoming power systems: 2

  • UPS type: Static UPS

  • Capacity: 10 minute runtime at full load

  • Redundancy: 2N (A+B)

  • Each server unit hosts two hot-swappable AC power supplies with 1 + 1 redundancy, auto-sensing and auto-switching capability

Backup power source

Data center is equipped with additional long term backup power source. Battery power provides power suply for up to 10 minutes. If both power lines are damaged, an additional on site power generator is used.

  • Generator Type: Diesel generator

  • Fuel Capacity: 24 hour runtime at full load

  • Redundancy: 2N

  • Fuel delivery: 24/7 via on-call contracts

Regular Data Center maintenance

These inspections are carried over by the data center collocation provider:

Frequency of activity

Activity

Frequency of activity

Activity

Monthly

Visual inspection and cleaning.

Quarterly

Technical tests of the following systems:

  • Access control

  • Security alarm

  • Fire detection and extinguishing

  • Cooling

  • Power supply


Cleaning and clearing of any growth outside the building.

Annually

Complete test of all technical installations, maintenance of buildings and surroundings.

For collocation services provided with double power feeds (A + B power), maintenance work is not performed on both power feeds simultaneously.
All maintenance activities are logged and documented.