Guest contributor |
Picture this: a live broadcast of the FIFA World Cup final is being transmitted across the globe using Microsoft Azure cloud networks. The video stream originates from South Africa, graphics are being layered in New York, and a voiceover is being recorded in Tokyo. All these elements need to be perfectly synchronised to ensure a seamless viewing experience. This is where Precision Time Protocol (PTP) comes into play, providing seamless synchronisation for these complicated, time-sensitive operations with lots of moving parts.
For the uninitiated, PTP is a protocol used to synchronise clocks throughout a computer network. It works by having a source clock, often referred to as the grandmaster, that is typically linked to a GPS signal. This connection to GPS provides highly accurate time references, down to the sub-microsecond level. This grandmaster clock distributes these time references across its network using switches and routers.
PTP is crucial for media, entertainment, finance and other high-performance applications to ensure synchronisation between data streams. Key standards and regulations like SMPTE 2110 for media, along with FINRA and MiFID II for finance, require the high-performance clock synchronisation that PTP provides. Beyond these other industry sectors like industrials, government, and service providers routinely rely on PTP for their production systems.
In larger networks or over long distances, PTP uses boundary clocks and transparent clocks to maintain synchronisation. Boundary clocks correct for any time discrepancies, known as ‘clock drift’, between the grandmaster clock and themselves, providing an accurate time source to the end systems. Transparent clocks, on the other hand, simply pass PTP messages through them, maintaining the integrity of the time information.
As industries that rely on precise time synchronisation migrate operations to the cloud, however, they’re hitting a barrier: most cloud providers don’t support multicast, a crucial component of PTP, or provide native precision time services. Multicast allows a single packet producer to move a stream of packets into a switch or router, which then replicates and distributes the packets to all subscribed entities. Without multicast and PTP together, these industries may face difficulties in extending their on-premises networks into the cloud, potentially slowing down plans to scale wider or enable remote work via the cloud or hybrid cloud setups.
Solutions like CloudSwXtch from SwXtch.io bridge the gap between traditional on-premises systems and the cloud. Within the Microsoft Azure Cloud universe, we can effortlessly add our CloudSwXtch virtual overlay network over the top of Azure technologies to deploy networks in scalable, reliable and expansive ways. This allows users to create new business opportunities with their existing network infrastructure, and multicast and PTP are both critical to enabling such opportunities.
Great things happen when CloudSwXtch and Azure come together. As Microsoft’s core networking is robust and well-designed, Azure technologies will reliably move packets without concern to underlying network performance while CloudSwXtch adds features on top of this core network. That means lower network congestion, higher reliability and simultaneous packet delivery – all key benefits of multicast distribution.
The benefit of simultaneous packet delivery is fortified through the timing and synchronisation capabilities of PTP, which SwXtch.io has also now enabled within CloudSwXtch through its partnership with Timebeat. CloudSwXtch integrates Timebeat’s PTP+Squared technology, enabling systems that rely on multicast – and by extension, PTP – to achieve the high level of synchronisation required to move to the cloud. When deploying CloudSwXtch, Azure tenants now have access to two valuable benefits for managing networking loads in the cloud – multicast and PTP – that were previously not available in public cloud networks like Azure.
CloudSwXtch also solves problems for Azure users new to these technologies. For example, the further that independent workloads are from the grandmaster clock’s time references, the greater the possibility for synchronisation errors. CloudSwXtch goes beyond just bridging PTP to the cloud; it also provides a solution to synchronisation problems for on-cloud workloads. By implementing two methods – straight PTP distribution and Timebeat PTP+Squared – CloudSwXtch ensures that virtual machines or microservices in the cloud can synchronise accurately.
This means that whether you're broadcasting a live event, executing high-frequency trades, or managing internet of things devices, you can ensure precise time synchronisation across all your systems and networks.
We have been working closely with Azure technologists to test, evaluate and verify that precise time synchronisation is assured. Azure allows applications like CloudSwXtch to access hardware timestamps from their Mellanox network interface controllers (NICs), which provides a higher level of time synchronisation accuracy. This can allow synchronisation of virtual machines running on Azure within a few microseconds of each other – more than sufficient for most applications.
With the technology now validated and available, chief financial officers looking to build scalable cloud infrastructure; chief technology officers and chief information officers seeking a broader range of products to deploy their systems; engineers and solutions architects creating solutions; and product teams developing new tools can all benefit from PTP in the cloud. Together, Microsoft Azure and SwXtch.io can offer our common customers improved operational efficiency, a seamless user experience and the ability to fully leverage the benefits of cloud technology.
Geeter Kyrazis is chief strategy officer at SwXtch.io
This article was originally published in the Autumn 2023 issue of Technology Record. To get future issues delivered directly to your inbox, sign up for a free subscription