
advertisement
Microsoft's online services started going down for users globally on 12 November. Outlook email and some associated workplace functions were affected for organisations that, in fact, had very little in common other than choosing Microsoft as their technology provider. In a preliminary report issued on 15 November, Microsoft identified three problems: a software bug introduced in July, increased traffic after Teams traffic was redirected to a system that also hosts Outlook, and fewer available servers.
There were good economic reasons for moving to shared systems. Earlier, organisations had to buy computers, house and maintain them, and hire specialists to keep them running. If each organisation has its own systems, then a failure in one is generally a failure for that organisation.
If thousands of organisations use a single provider, then a problem with that provider is a problem for all of them. A firm will calculate its own savings in choosing a cheaper, more convenient service. It will have no idea how many other hospitals, banks, or government agencies are making the same choice, and it will have no concept of the scale of the impact if that common supplier were to fail.
We can quantify that dependence. In Q3 2026, Amazon, Microsoft, and Google took 61 percent of the world's cloud infrastructure spend (at 26 percent, 19 percent, and 16 percent, respectively, according to Synergy Research Group). That is, of course, spent on rented computing infrastructure, not on the market for email and office software.
The two markets show how central the big providers have become to the day-to-day operations of other organisations. They can offer greater security and greater reliability through their scale, but they can also have a wider impact if they fail.
Different government departments may hire different software companies, but those companies could depend on the same technology provider for their computing needs. A failure at that provider could therefore disrupt several government services at once.
Take a bank keeping the copies of its database in multiple locations say in both Mumbai and Hyderabad. That sounds reassuring, but if both offices use the same system to check passwords, a failure there could prevent staff from accessing either copy.
Distance may help insulate the records from a problem at one site, but it provides little protection against a failure that affects both. An effective backup should therefore be independent of the system whose failure it is meant to protect against.
The bigger issue is how institutions assess risk. Each organisation will ask itself whether the technology it's buying is good enough for its own purposes. Very few will appreciate the extent to which the same underlying provider is present in other organisations' arrangements.
When customers all make sensible commercial decisions on their own, their aggregate dependence on that one provider can end up being far greater than any one individual customer would have anticipated.
Research supports the view that we should take this shared risk seriously. A Bank for International Settlements (BIS) study of over 100,000 cyber incidents found that cloud services were associated with lower costs for the smaller incidents. However, the same study found that depending on a provider that serves a large proportion of the economy would magnify the impact of a rare, catastrophic event.
The paper looked at the incidents rather than at the root cause of the Microsoft outage, but the warning about the degree of dependence is very much in play. A single organisation may be safer by using a specialist provider, even as the system as a whole becomes more exposed to a common failure.
Technology providers have customer contracts which contain agreed service levels. A contract may provide some form of compensation in the event of service interruption. Still, employees, users, other businesses, and public bodies with no direct relationship to the supplier will feel the impact. Their losses are generally not covered by the contract.
As shared systems become part of the economy, reliability becomes a matter of public and private oversight. That, in turn, means governments will have to decide which services are critical, what the service levels should be (and those levels will need to be independently testable), and what the collective risk is from the myriad organisations reliant on the same infrastructure.
That is a problem for traditional competition policy. Firms have increasingly opted for just a handful of successful providers as they offer lower prices and better products. A customer can, of course, find it very difficult to switch providers after spending years of effort working around their systems.
The Competition and Markets Authority (CMA), the UK's competition regulator, has stated that Amazon and Microsoft both have significant market power in the cloud computing market and has examined the barriers that prevent customers from using multiple suppliers. Enabling more freedom to mix suppliers can promote competition and give organisations greater choice in running their core services.
Reliance on shared platforms also raises questions about the responsibilities of the companies running them. Google's Firebase provides inexpensive tools for hosting websites and storing information, but fraudsters have exploited the same facilities to create fake banking pages and collect stolen data.
The Union Home Ministry identified misuse of Firebase in its 2023–24 annual report. In August 2026, Indian authorities sought removal of at least 57 Firebase-hosted websites and databases allegedly linked to financial scams. Preventing such misuse adds another responsibility for major technology providers alongside keeping their services reliable. They must make fraud harder without imposing excessive barriers on legitimate users.
These questions become even more pressing as India's digital services footprint continues to grow. As of September 2026, 2,330 departments were operating on the government's MeghRaj cloud (i.e. using shared computing facilities).This shows the growing reliance of government departments on shared computing facilities.
The Union Ministry of Electronics and Information Technology's March 2026 guidelines already classify government services by importance and call for backup arrangements for sensitive applications. Departments should therefore examine whether their alternatives can function independently, how quickly a service can be restored, and whether those arrangements have actually been tested.
India's policy response should highlight the common risks to all buyers, insist on independent recovery arrangements for core functions, and give regulators the data they need to assess exposure across institutions. Shared infrastructure has revolutionised access to technology. The efficiencies it has brought about now need to be matched by investment in the resilience of the economy as a whole.
(N.V. Geetha retired as Senior Technical Director at the National Informatics Centre after more than three decades in government service. She specialised in court informatics projects, designing and implementing IT systems for Indian courts and tribunals. This is an opinion piece and the views expressed above are the author’s own. The Quint neither endorses nor is responsible for them.)