3 ms·

AWS, or their own version of Box/Dropbox. (I guess that's why they have such a strong partnership with Box now).

derefr 11y ago
Working at (a company subsidiary to) IBM, I think there's a pretty unique culture of "buy rather than build" here. IBM has its own share of weird internal infrastructure (hello Notes/Domino), but for the things we don't "have our own version of", IBM seems perfectly happy to pick some small company and use their service, rather than entering a vertical just to have an answer for a competitor's entry. Basically, IBM is still fundamentally a consulting "solutions" company; their products are things they needed to build to solve a particular recurring consulting problem for their clients, that couldn't be solved with an application of off-the-shelf technology from a partner. This is reflected in IBM's own culture; we'll also look for an off-the-shelf solution[0]—or ask a partner to tailor their product to us—before even considering building something internal. "Scratching an itch" infrastructure projects like Facebook's Cassandra just don't really happen here. I doubt IBM could have "productized the infrastructure" the way Amazon did to end up with AWS; IBM doesn't have any unique infrastructure. BlueMix, such as it is, isn't a reflection of anything we were already doing; it was simply a de novo solution to a particular recurring consulting problem with unique constraints (high levels of isolation and on-premise collocation) that meant we couldn't just sell the client a solution by some IaaS partner of ours. For internal projects, we frequently still just use AWS or DigitalOcean or whathaveyou (although this is changing as BlueMix becomes more competent to serve our own particular needs as an "internal customer.") --- [0] ...as long as it's not built by one of our competitors in the "all things to all people megacorp" world. We use Slack, but turning on Google Docs integration is against security policy: we trust a small company like Slack with our IP, but trusting Google is right out.

Working at (a company subsidiary to) IBM, I think there's a pretty unique culture of "buy rather than build" here. IBM has its own share of weird internal infrastructure (hello Notes/Domino), but for the things we don't "have our own version of", IBM seems perfectly happy to pick some small company and use their service, rather than entering a vertical just to have an answer for a competitor's entry. Basically, IBM is still fundamentally a consulting "solutions" company; their products are things they needed to build to solve a particular recurring consulting problem for their clients, that couldn't be solved with an application of off-the-shelf technology from a partner. This is reflected in IBM's own culture; we'll also look for an off-the-shelf solution[0]—or ask a partner to tailor their product to us—before even considering building something internal. "Scratching an itch" infrastructure projects like Facebook's Cassandra just don't really happen here. I doubt IBM could have "productized the infrastructure" the way Amazon did to end up with AWS; IBM doesn't have any unique infrastructure. BlueMix, such as it is, isn't a reflection of anything we were already doing; it was simply a de novo solution to a particular recurring consulting problem with unique constraints (high levels of isolation and on-premise collocation) that meant we couldn't just sell the client a solution by some IaaS partner of ours. For internal projects, we frequently still just use AWS or DigitalOcean or whathaveyou (although this is changing as BlueMix becomes more competent to serve our own particular needs as an "internal customer.") --- [0] ...as long as it's not built by one of our competitors in the "all things to all people megacorp" world. We use Slack, but turning on Google Docs integration is against security policy: we trust a small company like Slack with our IP, but trusting Google is right out.