Microsoft Hailstorm

» see microsoft.com/myservices as it lived, on the wayback machine

~ the obituary ~

hailstorm sat in the part of internet history i can read about but cannot personally remember as a user. it was never a consumer product. the only thing anyone ever saw of hailstorm was a bill gates demo and a year of trade-press speculation about whether microsoft was going to own everyone's contact list.

the pitch on march 19 2001 was that microsoft would run a set of soap web services holding your identity, contacts, calendar, mail, notifications, location and a few other slices of personal data. third parties would call into these services with your permission, and the whole thing would be authenticated by microsoft passport. it was the most ambitious thing the.net strategy had publicly attached itself to. bill gates personally previewed it at a redmond press event, and it was promised at the professional developers conference later that year.

by the fall of 2001 microsoft had renamed it from hailstorm to the less aggressive.net my services. by april 11 2002 the company admitted to the new york times that no third-party providers had signed up. without partners willing to plug their own customer data into a microsoft-hosted broker, there was no product. the news got reported by the register and slashdot the same week, and the brand was retired without a closing ceremony.

~ the case file ~

~ the announcement ~

the march 2001 launch was the kind of microsoft event that does not happen anymore. gates appeared in person, the press release ran to several pages, and the framing was that hailstorm would be the connective tissue for the entire next decade of personal computing. the technical pieces were genuine. soap, uddi and ws-discovery were real protocols microsoft was helping define, and the idea that you would have a remote home for your calendar that any app could read with your consent did eventually become a normal thing twenty years later.

the bet was that microsoft could be the only company holding that data. that was where partners balked. banks were not going to route account info through passport, and telcos were not going to hand microsoft a hook into their subscriber records. credit-card processors took the same position about their payments stack.

~ the actual end ~

there was no shutdown notice because there was no service to shut down. what microsoft killed in april 2002 was the consumer ambition of the project. the.net my services name stayed alive as a developer-tier offering for a while, and pieces of the technology got absorbed into windows server and what later became the identity stack inside microsoft online services. the consumer hailstorm vision, of a microsoft-owned vault of your personal information, was dead by the time the trade press was finished writing obituaries.

the register's coverage is the cleanest period record of why it fell apart. their reporting at the time noted that consumers had not even gotten the chance to object, because banks and content owners and large telcos had already objected first. the centralized-data-broker model needed permission from the institutions that owned the data flows, and those institutions had no interest in becoming dependent on a microsoft-run middleman.

~ the next chapter ~

hailstorm is one of the most studied non-shipping products in microsoft history. it gets cited in privacy-policy papers and in identity-protocol histories whenever someone explains why federated authentication became a standards process rather than a single-vendor product. the openid effort that started a few years later was partly a reaction to hailstorm. so was the entire saml-and-oauth identity stack that runs the modern web.

microsoft itself returned to almost the same vision a decade later under different branding. windows live and the later microsoft account brand are direct descendants of the same impulse, with the modern graph api playing the developer-facing role hailstorm had wanted. by the time those products shipped the world had decided which side of the broker-vs-protocol argument it preferred, and hailstorm is part of the reason that decision was already made.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »