We live in a world of highly connected computer networks, with digital highways forming the very essence of our economic and social structure. We do our banking online, respond to customers and potential business prospects over the Internet and we live our lives immersed in the realm of social media. Every single microsecond of the day is filled with digital transactions that echo across the planet and bring our ultra-connected world to life.
The first study of the science of network reliability was published in 1956 by C.E. Shannon and E.F. Moore. From there, networks have evolved to be highly sophisticated. Network reliability makes our modern lives thrive, and our connections and transactions are made possible because of the robustness and resilience of our digital networks. We rely on these connections to power our society, create wealth, enhance our social lives, and bring us entertainment with a simple touch of the screen.
Creating robust, reliable large-scale networks is both an art and a science. You will need to assess the needs of your network, and carefully plan for expected peak traffic loads, changes and possible failures. Sophisticated network testing and monitoring tools must be in place to assure continuous uptime. Network redundancy should be boosted by adding extra paths to the network, and network security best practices put in place. These security safeguards would include encrypting network traffic, implementing anti-virus software and firewalls, and using strong authentication methods.
Several important metrics can be used to measure network reliability and performance. Network uptime is the percentage of time in a year the network is up and running. Latency measures how long it takes for a data packet to travel between points on a network. Jitter refers to an inconsistent arrival of data packets. If jitter is above 20ms (milliseconds), or latency above 150ms, you can expect delays in audio or video. Packet loss, another key metric, occurs whenever a packet arrives out of order, too late or does not arrive at all.
With our reliance on computer networks, it’s not surprising that network outages make international news. In January 2023, Microsoft experienced a global disruption that affected many of their online services. February saw a Microsoft Outlook international service outage, while in March, Reddit experienced a disruption affecting their global users.
Payment processor Square had system-connectivity issues and an outage that prevented their merchants from being able to conduct business and process transactions on September 8th. Two months later, Australia’s Optus network suffered 100% packet loss on both their mobile and broadband networks across the country, affecting over 10 million people.
Government-induced internet shutdowns have been used to curtail freedom in many countries. Unfortunately, these restrictions have also resulted in an €8.2 billion loss to the world economy.
The longest internet outage in Manipur, India, lasted for over 5,000 hours. Iran had 66 different outages during the year. Restrictions in Russia had the biggest economic effect, costing the country’s economy €3.7 billion.
The Internet is everywhere, and computer networks make all of our banking, communications, entertainment, travel and remote work possible. The importance and omnipresence of the internet is perfectly illustrated if you stop to consider how much a day without the internet would cost.
According to information from a leading international security firm, the economic impact of one day without internet across the globe would be €39.5 billion. Seven countries would each suffer a loss of over €1 billion in that single day.
When it comes to network design, there are several topologies that can help ensure reliable operation. A topology is the logical or physical arrangement of nodes and connections within a computer network. In a star configuration, every node is connected to a central switch. With mesh topology, the nodes are interconnected using point-to-point connections. This interconnectivity greatly reduces the chance of failure. Hybrid networks, which are composed of two or more different topologies, offer greater flexibility but may also be more difficult to manage.
One of the keys to ensuring network availability is creating a robust infrastructure built on proven, reliable hardware. Businesses that fail to invest in needed infrastructure will suffer financially due to lost productivity, and may open themselves up to risks due to data breaches and data management issues. Software solutions that help manage the health of the network can deliver valuable insights into network performance and can also help companies pre-empt disruptions.
In order to ensure data integrity and information delivery, critical network protocols are used. Data transmission is standardised, as networking protocols are used to define data routing, error handling, security and device interoperability. Network operators also employ sophisticated practices for securing networks against attacks and breaches. These may include network segmentation and segregation, effective firewall positioning, restricting physical access to data centres and the use of intrusion protection and prevention systems.
Every company and enterprise needs to have a proven plan and strategy for rapid recovery from network or system failure in order to ensure continuity of their business. Downtime and loss of data and operational capacity may result from malware or ransomware attacks, faulty hardware or software, unauthorised access or security issues and natural disasters.
With a proven and tested disaster recovery strategy, a company can dramatically reduce data loss or disruption. In addition to conducting thorough reviews of your plans, you should run disaster test drills that turn your plan and specific recovery scenarios and procedures into action.
As of December 2023, there were approximately 11,000 data centres in the world. The United States accounted for over 5,000 of those centres, with Germany and the U.K. totaling over 500 centre locations each. Data centres house storage devices, servers and network routers, and help assure safe, continual access to applications and data.
When constructing a data centre, operators pay particular attention to key design principles like modularity, energy efficiency and scalability. Energy efficiency is measured using statistics like power usage effectiveness (PUE), which is calculated by dividing the amount of power coming into a facility by the power that is used to run the IT equipment. As digital demands rise, large data centre operators want their designs to be scalable and replicable when expanding to additional locations. New options like PFMs, or prefabricated modular data centres, allow operators to expand to demand as needed.
As of December 2023, there were approximately 11,000 data centres in the world. The United States accounted for over 5,000 of those centres, with Germany and the U.K. totaling over 500 centre locations each. Data centres house storage devices, servers and network routers, and help assure safe, continual access to applications and data.
When constructing a data centre, operators pay particular attention to key design principles like modularity, energy efficiency and scalability. Energy efficiency is measured using statistics like power usage effectiveness (PUE), which is calculated by dividing the amount of power coming into a facility by the power that is used to run the IT equipment. As digital demands rise, large data centre operators want their designs to be scalable and replicable when expanding to additional locations. New options like PFMs, or prefabricated modular data centres, allow operators to expand to demand as needed.
