Coinbase Unveils AI Solution to Predict Traffic and Scale Databases

Coinbase announced on 26 August that it has successfully developed and deployed a new machine learning model capable of predicting spikes in user traffic and automatically scaling its resources to prevent downtime and enhance platform efficiency.

This AI solution is designed to address the recurring issue of platform crashes during unpredictable traffic surges, which have been a significant challenge during volatile market conditions.

By leveraging advanced analytics, Coinbase can now preemptively scale its resources, effectively mitigating the risks of system outages and significantly improving operational efficiency.

The blog details:

"Starting to scale when traffic is already high is often too late. We developed an automatic scaling solution that uses machine learning to predict the traffic spikes and trigger a scale up before the traffic arrives.”

The exchange has reported that the model has already demonstrated its value during recent periods of market volatility.

Coinbase noted:

"As traffic increased so did our scale target, doubling twice a few hours before peak traffic. The model continued scaling up and down with the daily usage pattern, until volatility decreased and there was no longer a need to scale up."

The model's primary task is to provide a signal with a 60-minute lead time before a traffic spike occurs.

Initially, Coinbase experimented with a time-series forecasting model that aimed to predict traffic levels 60 minutes into the future.

However, this approach was found to be ineffective due to the lag time in the underlying statistics.

To overcome this, Coinbase transformed the problem into a longer-term classification challenge.

The new model utilises external signals such as price fluctuations in major cryptocurrencies like Bitcoin and Ethereum.

It aims to determine whether traffic will exceed a certain threshold level in the next few hours.

This approach has significantly improved accuracy compared to previous methods.

The blog states:

"The key insight: if cryptocurrency price volatility is high and the current traffic is approaching the target level at a faster rate than anticipated, then the likelihood of a traffic spike is increased."

Coinbase explains that its AI model was carefully designed to strike a balance between avoiding missed spikes and minimising false alerts.

This balance is crucial because triggering false alerts can lead to wasted resources, while too few false alerts might result in an inaccurate model that fails to prevent platform crashes during high-traffic periods.

Coinbase's History of Outages

Coinbase has faced challenges with market volatility, experiencing notable outages and technical issues at critical moments in the cryptocurrency market.

On 14 May 2024, the exchange suffered a major outage that lasted over three hours, affecting both its desktop and mobile platforms.

Users encountered a '503 service temporarily unavailable' message on the website and incorrect 'planned maintenance' notifications on the mobile app.

Just two months earlier, as Bitcoin was nearing its all-time high, Coinbase crashed, leaving thousands unable to capitalise on their gains.

A Coinbase spokesperson commented:

"We are encouraged by the excitement in the market and are continuing to enhance the capacity and resilience of our systems."

Moreover, one month prior to that incident, Coinbase experienced a crash due to its inability to manage unexpected traffic during a bull run.

Users have also reported financial losses stemming from the inability to execute trades during these outages, with Coinbase advising users to refresh the site until services are restored.

Coinbase Striving to Improve Platform's Resilience

In response to previous difficulties, Coinbase has vowed to strengthen its platform's resilience.

Central to this effort are investments in server capacity and software optimisation.

The company has repeatedly emphasized its commitment to resolving past issues, with a strong focus on technical enhancements and infrastructure improvements.

Coinbase has dedicated resources to expanding its server capacity and refining its software architecture to better withstand surges in traffic.

Yet, recognising that the challenges stem from abrupt traffic spikes rather than gradual increases, the application of AI to forecast these anomalies and dynamically scale databases seems to be the most promising strategy—provided it proves effective.

The performance of this technology will be evaluated during future episodes of market volatility, potentially offering a glimpse into the advancement of technology-driven solutions in cryptocurrency trading.