Cloud Chaos: UniSuper's Recovery Stumbles After Google Cloud Mishap, Exposing Industry-Wide Vulnerabilities

$125bn AUM Fund Manager UniSuper Faces Prolonged Service Disruption After Google Cloud Account Deletion

UniSuper, an Australian fund manager boasting assets under management (AUM) of approximately $125 billion, has encountered significant disruptions in digital services due to an erroneous removal of their Google Cloud (GCP) account. Despite ongoing efforts, the restoration has been incomplete for more than ten days since the incident occurred just before May 2nd.

The CEOs of UniSuper and Google Cloud, Thomas Kurian, released a joint statement explaining the cause as "an inadvertent misconfiguration during the provisioning of UniSuper's Private Cloud Service." This error led to the accidental deletion of UniSuper's cloud subscription, with Google accepting full responsibility for the mishap. Fortunately, UniSuper had preemptively secured their data on an alternate infrastructure, preventing a total loss. However, the recovery time, extending beyond ten days, underscores the severity of the disruption.

Currently, UniSuper members are able to log into their accounts and view their balances, although the displayed information may be incomplete or inaccurate, as it does not include transactions pending due to the outage. While transaction initiation remains possible both online and over the phone, processing delays are expected. This poses significant inconvenience for members needing immediate financial actions like investment changes or withdrawals.

Five months prior to this incident, Angelo Furina, UniSuper’s Head of Enterprise Infrastructure & Cloud, celebrated the successful integration of GCP during a podcast. He initially overlooked GCP due to concerns over its readiness for enterprise-level demands but eventually selected it for its cost efficiency and the flexibility of its VMWare solutions. This decision was aimed at accommodating the rapid integration needs following UniSuper's active mergers and acquisitions. Despite these initially perceived advantages, the recent issues have highlighted potential drawbacks.

Analysis

In standard cloud transformations, infrastructure provisioning is automated—often described as "infrastructure as code." This automation typically allows for quick deployment and management of IT resources, which should not necessitate a recovery time as long as ten days. The extended downtime at UniSuper suggests potential deficiencies in their IT practices or perhaps a misalignment among various departments, indicating room for improvement in coordination and efficiency.

Market Impact on Google Cloud

This incident not only affects UniSuper but also poses a reputational risk for Google Cloud, which is already a smaller contender in the enterprise computing market, dominated by AWS and Azure. Despite GCP's contribution of approximately 12% to Alphabet’s total revenue and a substantial increase in operating income recently, the ongoing service issues could have long-term negative effects on customer trust and financial performance. At the last quarter end, Google Cloud stood out among Alphabet's business units with a remarkable 371% increase in operating income. Given Wall Street's recent profitability forecasts, it's fair to say that Google Cloud played a pivotal role in helping Alphabet meet its financial targets. However, the latest incident with UniSuper has raised concerns among both existing and potential Google Cloud customers. Questions about the reliability of the service are mounting: How many hidden flaws remain? When might these issues affect my organization? These ongoing concerns are likely to influence Google's financial performance negatively over time.

Competition in the Public Cloud Sector

The public cloud industry, often compared to utility sectors like electricity and gas, is experiencing intense competition. Providers are struggling to attract new clients, particularly large organizations, because they frequently fail to meet the key expectations of chief information officers, such as reasonable pricing, flexibility, and robust security and compliance. This intense competition has led to a vicious cycle where cloud providers are compelled to increase their marketing and sales budgets to stand out, which in turn may lead to higher prices, pushing them further away from the interests of large enterprise buyers. For instance, it is reported that AWS allocates more than 30% of its total spending to marketing and sales, a figure that has been rising annually. A senior manager, a major enterprise customer of Azure, told CTOL.digital, "There is reluctance (from the cloud providers) to cut marketing and sales budgets, yet in every meeting, there are always several high-paid individuals who contribute little to the conversation, earning between $300,000 and $500,000."

Did you know?

Recent Outages Across Major Public Cloud Providers

• AWS: Experienced a three-hour major outage in June 2023 affecting 104 services including popular platforms like Fortnite and McDonald's.
• GCP: A significant downtime in April 2023 took down an entire region for about a day, with extended disruptions lasting two weeks.
• Azure: Faced an eight-hour partial service disruption in July 2023 due to severe weather in the Netherlands.