Comment by manjose2018

13 hours ago

Does your company have a ton of thick-client apps? I feel like so many workloads are on the web right now and thick clients are slowly becoming a thing of the past - but maybe in Enterprise there are still stubborn thick client Win32 apps that are mission critical.

If I were starting a company today I would make sure that the majority of my apps were web based. (CRM[netsuite or sf.com], O365/GoogleDocs, workday|gusto, etc.)

5 comments

manjose2018

Reply
kjs3  1 hour ago

At my last enterprise gig (financial, ~85k employees), when I inventoried apps for cloud migration, there were around 1500 what I guess you'd call 'thick apps', and about 75% of them were some variation of "stubborn thick client Win32 (or .NET) apps that are mission critical". Mission critical to someone, anyway, and the number of apps with 1 or 2 users was kinda eye opening. Noone was going to fund the effort to rewrite them to be cloud native or web based, so there was a lot of lifting and shifting going on.

Many businesses might do better pushing everything to web apps, but SaaS introduces a whole bunch of other issues in highly regulated worlds like finance. It's not always the right answer.

d3Xt3r  12 hours ago

Most high-security environments (banks, govt agencies etc) have a large on-prem footprint. Then there's specialised industries like manufacturing, science and engineering who typically have a bunch of embedded systems, often running ancient versions of Windows or something that needs to talk to a local ancient server. And many old companies have tons of legacy apps which are a PITA to migrate to newer infrastructure. So yeah, plenty of folks out there using omprem gear, it's kinda the reason why Broadcom is able to what it's doing - lots of suckers out there who're married yo vmware and can't get a divorce.

  • kjs3  27 minutes ago

    Couple of things (having lived in this world). The people running this stuff aren't by default stupid (well, not always), and not assuming they are helps with understanding.

    First, someone has to pay to migrate apps, and as often as not the resources to do that aren't available. Sometimes you do the math and it doesn't make finacial sense to allocate those resources. And sometimes you just can't move apps. The vendor might not support vm/cloud/web deployment. Or, very common in manufacturing, you have an app that runs a machine and requires a very, very specific hw/sw environment to be supported.

    Next, there's that whole regulated thing. We spend a lot of time in front of federal/national regulators explaining how we do things so that we don't screw up your mortgage or whatever. And those regulators take careful notes. So if I've spent the last 10 years explaining how we have this amazing VMWare deployment that ticks all the boxes, they're going to want some very, very detailed explanations of how HP VME or Proxmox or whatever is going to tick all the same boxes at least as good as VMWare did. And we're inviting a lot of scrutiny for a couple of years to make sure we did it right. It's a lot of work.

    Sometimes you weigh everything, not just 'it costs less' or 'Broadcom are a bunch of assholes' or 'we don't want bros on HN to think we're suckers', and the answer is legitimately "the divorce isn't worth it".