How important is resilient time synchronisation when choosing a data centre?

Author: Carly Weller

By Richard Hoptroff, CTO, Hoptroff 

Previously, precision timing has not been a necessity for data centres. Far from standard practice, if a customer needs accurate time synchronisation, they buy it in. However, this data centre hesitancy regarding new timing solutions such as Traceable Time as a Service (TTaaS) cannot be sustained in the face of growing industry demands for accurate time by data centres to improve the quality of operations.

Data centre customers are progressively distributing computational processes in the cloud. While this transition can offer greater efficiency and innovation it also presents dangers such as fragmentation which can result in compromising data quality and integrity. A casualty of this process is the coherence of time, and with it the ability to establish causality between events, especially when relying on traditional Network Time Protocol (NTP) time feeds that offer less accuracy and verifiability than a software-based Precision Time Protocol (PTP) solution. As the speed of data transfer and cloud migration increase, the need for accuracy and traceability in time also increases in data management. Data centres will experience growing demands for highly accurate, traceable, and resilient time from their customers.

Accurate, resilient, and traceable time is becoming increasingly important for data centre operations. This is particularly true when providing sub-microsecond delivery in data centres. Many computers must interact and have clock synchronisation to regulatory required speeds. Within the finance sector, the MiFID II and Consolidated Audit Trail (CAT) regulations require companies to have trusted timestamping on transactions to as much as a hundred millionths of a second (100µs). Thus, they need a reliable source of time that is precise and verifiable to ensure that transactions are compliant. This is not, however, just a compliance issue. If timestamps are wrong, counterparties will reject the trades and business will be lost. 

Within the media and broadcasting sector, companies can cut costs considerably if they send content over Internet Protocol (IP) rather than the more traditional Serial Digital Interface (SDI). This also enables them to process content in the cloud rather than on-location in expensive edit suites. The catch is that IP is asynchronous, so the content needs to be resynchronised at every point along the production chain. This is where a source of traceable and resilient time is needed, to ensure that this content is synchronised correctly at whatever point in the chain it is in.

Traceable time stamps improving data quality

This gradual shift to virtual systems was accelerated by the pandemic. Almost all businesses now have some dependency on the cloud, from cashless payments to appointment booking. Remote working is also becoming more significant and companies that use data centres need to ensure that their data is secure in the cloud. Traceable and accurate time is necessary now more than ever to make sure that documents can be worked on in the cloud by multiple people at the same time to maintain business efficiency.

As the accurate recording of data becomes more central to compliance and regulation, so does trust. Companies need to be able to ensure that their time is correct and be able to prove it to their customers. This scenario is even more important when virtual events have real-world effects where someone needs to be held accountable. Blockchain, data validation, and GDPR regulation are some of many use-cases that require the level of certainty that traceable timing solutions provide.

Real-time regulation, monitoring and reporting

Accurate time is traditionally achieved with Global Navigational Satellite Systems such as GPS, in combination with a grandmaster clock that translates this into the PTP internet format used to transmit time to servers in a data centre. As we push towards processing data at the edge, more and more equipment setups are required, and the cost of this approach soon becomes prohibitive. It’s not only expensive, but very high maintenance and typically implemented by a network engineer who is not a timing specialist. The only way to be certain of the time is to source it from three independent sources (e.g., GPS, Glonass and Galileo) and use three grandmaster clocks – but this makes it three times as expensive. The final issue that has been imposed by COVID-19 is the inability to visit the data centres to maintain all this equipment. 

With the traditional methods used to achieve traceable time, the challenge lies with the satellites themselves. Jamming and spoofing attacks are becoming increasingly common and pose a real threat to these systems. Intentional interference (‘jamming’) is increasingly common, most anecdotally when commercial drivers deliberately jam their on-board tracking devices to cover their tracks. Spoofing is a more sophisticated form of interference where false GNSS signals are generated to convince GNSS receivers that they are somewhere else. Originally developed by state actors as a defensive strategy, they are now available at low cost and will increasingly be used as a defence against undesired drone incursions as drone technology becomes more prevalent. 

Resilient and reliable timing solutions

When dealing with satellite interference, network-derived timing sources are the solution. These timing sources allow a software agent to synchronise the server system clock; and the monitoring, alerting, and logging system ensures that time synchronisation is maintained. It is a more precise, and now verified, timing environment that enables global financial services and other industries to enhance the value of their data, optimise business operations and efficiency. There is no hardware involved, it is a software solution.    

Methods like Traceable Time as a Service offer lower costs and removes the need for cumbersome physical hardware installation. With increasing reliance on the cloud, having a terrestrial time distribution method increases security even in the event of satellite signal loss. Clock synchronisation is a simple compliance task which is not the institution’s focus, so out-sourcing it can save time and hassle. 

As more traceable timing solutions launch due to the increasing capacity of data centres across financial services and beyond, companies will look to invest in this invisible utility. Players in the data centre sector, such as Equinix, are catching up with this market trend and has recently launched its version of ‘Time as a Service’. This emphasises the importance of having traceable and accurate time throughout data centres, while ensuring that customer’s data is safe. It is likely we will see demand grow further as more systems move to the cloud, so businesses should look to an experienced timing solution provider to guarantee that their time is precise, accurate and verifiable.