Proposed server purchase for GitLab.com

Re:Proposed server purchase for GitLab.com by lykron on Sunday December 11, 2016 @05:27PM (#13153762)

With regards to the switches, I would argue that they should skip SDN switches all together and get some Cisco Catalysts as TOR Switches. 2x 48-port switches for each rack with redundant core routers in spine-leaf. SDN is cool, but at the current scale, it seems like it would be more resource intensive than it is worth.

[ Parent | on HN ]

Re:Proposed server purchase for GitLab.com by wmf on Sunday December 11, 2016 @05:33PM (#13153794)

This is one rack of equipment; no spines are needed. http://blog.ipspace.net/2014/10/all-you-need-are-two-top-of-...

Cumulus isn't really SDN so I'm not sure what you're saying there.

Traditional networking is fine but it's totally different than Linux so you need dedicated netops people to manage it. And Cisco is the most expensive traditional vendor.

[ Parent | on HN ]

Re:Proposed server purchase for GitLab.com by lykron on Sunday December 11, 2016 @05:39PM (#13153830)

I'm not super well-versed in the networking area. I misread the posting. I thought they were looking at 64Us of servers.

I've just gotten the notion that SDN is almost ready for primetime, just not yet.

[ Parent | on HN ]

Re:Proposed server purchase for GitLab.com by bronson on Sunday December 11, 2016 @06:38PM (#13154230)

It's been almost ready for primetime for over a decade now.

[ Parent | on HN ]

Re:Proposed server purchase for GitLab.com by mohctp on Sunday December 11, 2016 @10:04PM (#13155184)

Agreed, skip SDN. Skip MC-LAG.

Juniper, Cisco, and Arista all have solutions for this environment.

[ Parent | on HN ]