May 2017 - Routing & Reachability | Latency

Learn From Einstein And Synchronize the Internet

During Einstein’s days in Bern, the idea that all clocks should tell the same time was a novelty. Their synchronization went hand in hand with changes that altered the world. What does this tell us about latency in our current Internet infrastructure? Mathias Röckel explains.


© GoShiva | iStockphoto

When young Albert Einstein worked at the Swiss patent office in Bern, the city featured a variety of clocks the people had every right to be proud of. These clocks performed their duty with the highest degree of accuracy – and they were synchronized. They all showed the same time without any measurable discrepancy. This was no gadgetry.

Railroads and telegraphy had just changed the face of the earth. Geographic distances were rapidly losing their relevance. Time was becoming more important. Getting things done at the anticipated time was becoming more important. The world was becoming smaller and it ticked faster. People were less tolerant towards asynchronism. Investing in infrastructure necessary to measure time was considered an important public interest – and supported accordingly.

Today, a watch synchronized with an atomic clock via satellite costs less than a full tank of gas, but asynchronism still gives us headaches. Take the stock market, for instance.

A stock’s price isn’t as dependent as some might think on how well the company is doing and whether its gains are reasonable. Much more important is what we believe about the company’s future (or what we believe others believe). Stock prices are detached from the real economy, and we are used to that.

What we refuse to get used to are traders who don’t care about the future of the companies they trade. They don’t care if a stock is going up or down. Their only concern is speed and their business having a millisecond advantage that traders can use to place orders right before a stock’s price changes – and that they know in which direction the price moves.

Taking advantage of those who miss real time by milliseconds

That High Frequency Trading is even a business displeases many. The reasons are understandable. To win a game by beating those who miss real time by milliseconds is hard to classify as fair play. It is also against macro economic reason. What makes High Frequency Trading harder to digest is that winning has become a matter of access to the right technology. Even worse, flash crash scenarios have given the world a new fright that leaves experts clueless. This is not how Homo oeconomicus pictured globalization.
It is laudable if national governments try to hedge such developments. Regulation and transparency, if applied with a sense of proportion, are seldom a mistake. But in the connected digital world, this kind of legal regulation works much like taxation and tax shelters. Money and data streams travel the path of least resistance. But a global political solution to the issue is not in sight.

Instead of calling for more regulation, why not trust in technology instead? And by trust, I mean: invest. It has worked before, as the history of DE-CIX proves. When the Internet Exchange was – privately – founded more than 20 years ago, it was a given that IP-based data was routed via North America. This resulted in higher than necessary latency, an issue the industry tackled back then, but is far from having resolved: The time it takes to send and receive data is the single fundamental issue responsible for our asychnronized networks. 

In hindsight, it was the right decision to get providers to work together and connect them with one another. Nobody will dispute this today. The digital infrastructure has become as important for the functioning of our society and economy as, for instance, the airport or the railway. Back then, however, it took a leap of faith.

The Internet is far from being completed. The data flood is ever rising. Most Internet users have several devices. The Internet of Things is still in its infancy. Nobody knows where it will end. You don’t have to be Einstein to reach the conclusion that we cannot be content with the current digital infrastructure. That includes latency and is not limited to those trading in stocks. Or cloud computing. Or industry 4.0. Those with better connections have the upper hand over everybody else.

This is why we need to seriously invest to make our digital infrastructure future proof. We need commercial providers who place their bets on the free flow and exchange of data. We need frameworks that encourage investments that will serve the common good, a common good that does not end at the borders of our domestic markets. We should work toward a digital infrastructure that offers equal opportunities without discrimination of users, services, or data. Real time and the speed of light should be the target. We should try to give everybody high quality access to the Internet so all can access the same information and participate in the markets based on the same prerequisites and with the same user experience. Only then will we live, to use a popular phrase from the last millennium that has unfortunately been forgotten lately, in a global village.