Plex Cloud

» see plex.tv as it lived, on the wayback machine

~ the eulogy ~

Plex Cloud went into invite-only beta on 26 September 2016 and solved the one complaint every Plex user had. A Plex server needed a machine that stayed on, which in practice meant an old desktop humming under a table, or a NAS, or a laptop nobody was allowed to close. Plex Cloud moved that machine into Plex's own infrastructure and pointed it at storage you were already renting.

The launch partner was Amazon Drive, which at the time sold unlimited storage for about $60 a year. You needed a Plex Pass on top of that, $4.99 monthly or $149.99 once. For roughly the price of a hard drive per year you got a media server that could not be unplugged by somebody wanting the socket.

It worked the way a demo works. Then it met the part of the arrangement Plex did not control.

~ the dossier ~

~ the amazon problem ~

Within about two months users could not reliably reach their own files. Threads filled up with people wondering whether Amazon was throttling, rate limiting, or simply refusing the access pattern, because a media server hammering a consumer storage account with random reads looks nothing like the backup workload Amazon had priced for. TechCrunch wrote it up on 2 December 2016 under a headline saying Amazon was not playing nice.

Plex responded the same day by opening the beta to three more storage providers, Dropbox among them, which is what you do when your only partner has stopped answering. Amazon killed its unlimited tier the following June anyway. The product's founding economics had a nine month life.

~ twenty months of trying ~

Plex Cloud opened to every Plex Pass subscriber on 9 March 2017 and that was the high point. In February 2018 the company quietly stopped letting anybody create a new cloud server, which is the move that tells you a decision has already been made upstairs. The announcement came on 11 September 2018 and the servers went off on 30 November.

The wording was unusually plain for a shutdown notice. Plex said it had been unable to find a solution capable of delivering a first class experience at a reasonable cost, and that it needed the engineering time for core work. Running somebody else's transcoding job against somebody else's storage bill is a business with no good version at $5 a month.

None of this was ever going to be a Turkish product anyway. A home media server here was a box full of files you had already downloaded, and the upload speed on a domestic line made the idea of pushing that library into anyone's cloud a several-week proposition. Plex Cloud assumed a fat pipe going the wrong direction.

~ what it was testing ~

The idea keeps coming back because it is correct. Personal media on rented infrastructure is now how most people store everything, and Plex itself eventually went the other way entirely, into ad-supported free streaming, which is where the money turned out to be.

What died in 2018 was a specific bet: that consumer cloud storage was cheap enough and open enough to be treated as a disk. It was cheap. It was never open.

@}-->--

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

« previous  ·  random  ·  next »