7 ms·

The amount of throughput you can get out of AMD EPYC zen5 servers for the price of a basic mainframe is insane. Even if IBM wins in single core perf using absur

The amount of throughput you can get out of AMD EPYC zen5 servers for the price of a basic mainframe is insane. Even if IBM wins in single core perf using absurd amount of cache and absurd cooling solution, the total rack throughput is definitely won by "commodity" hardware.

rbanffy 2y ago
Maybe, but then you need to engineer the 99.99999% uptime yourself. If it were actually cheaper, IBM wouldn’t be selling these machines so well.
- oldpersonintx 2y ago
  [dead]
- wolf550e 2y ago
  I don't think they are actually selling those machines so well. They have a captive legacy customer base, who else is buying those?
  - neverartful 2y ago
    From TFA: "Overall, Z is growing very healthily. LinuxONE is the fastest area of growth for us right now." However, he didn't elaborate or give any examples. If I were the interviewer, I would have followed it with: "Oh?! Can you provide some examples for the readers who believe that you only sell to captive audiences?"
    - rbanffy 2y ago
      If I had the need to ridiculously scale up a Linux workload, I’d immediately think LinuxONE or POWER. If I had the R&D budget for an on-prem VM hosting box, I’d seriously consider their smaller Express 4, which is not much pricier than a similarly capable x86 machine.
    - IanCutress 2y ago
      Because I've covered that in a previous video :) https://www.youtube.com/watch?v=ZVO55oxJ464 https://www.youtube.com/watch?v=ZVO55oxJ464
- Muromec 2y ago
  They are mostly selling to the captive audience who is 40 years deep into COBOL and can't pull out until it falls on top of them.
  - feurio 2y ago
    Well, you know what to do - just sell a system that runs COBOL on an x86 cluster or Kubernetes or whatever, and you'll make billions! Be sure to allocate me a bunch of shares for giving you the idea.
  - pjmlp 2y ago
    Many of those machines are running Java workloads, also COBOL isn't the only mainframe language, and the best thing in terms of security is that they don't use C as systems language, rather saner stuff with proper arrays, strings and bounds checking. That is why Unisys ClearPath MCP is still a thing, tracing back to its Burroughs 1961 heritage, security above all.
    - rbanffy 2y ago
      Java workloads might be easier to port away from but aren’t MCP machines x86-based nowadays? Unisys did a ton of work for Microsoft making itself obsolete by adding a lot to their secret multiprocessor stuff to Windows. It’s amazing how Microsoft convinced so many companies to shoot their feet.
      - pjmlp 2y ago
        They still use NEWP as systems language, there is no C stuff to exploit.
        rbanffy 2y ago
        I had a lot of trouble with the idiosyncratic ways of MCP, and I have no doubt Tron's villain was named after it.
  - rbanffy 2y ago
neverartful 2y ago
These comments always come up with every mainframe post. It's not only about performance. If it were it would be x86 or pSystems (AIX/POWER). The reason customers buy mainframes is RAS (reliabililty, availability, scalability). Notice that performance is not part of RAS.
- jiggawatts 2y ago
  You and the parent are both "missing the point", which is sadly not talked about by the manufacturer either (IBM). I used to work for Citrix, which is "software that turns Windows into a mainframe OS". Basically, you get remote thin terminals the same as you would with an IBM mainframe, but instead of showing you green text you get a Windows desktop. Citrix used to sell this as a "cost saving" solution that inevitably would cost 2-3x the same as traditional desktops. The real benefit for both IBM mainframes and Citrix is: latency. You can't avoid the speed of light, but centralising data and compute into "one box" or as close as you can get it (one rack, one data centre, etc...) provides enormous benefits to most kinds of applications. If you have some complex business workflow that needs to talk to dozens of tables in multiple logical databases, then having all of that unfold in a single mainframe will be faster than if it has to bounce around a network in a "modern" architecture. In real enterprise environments (i.e.: not a FAANG) any traffic that has to traverse between servers will typically use 10 Gbps NICs at best (not 100 Gbps!), have no topology optimisation of any kind, and flow through at a minimum one load balancer, one firewall, one router, and multiple switches. Within a mainframe you might have low double-digit microsecond latencies between processes or LPARs, across an enterprise network between services and independent servers its not unusual to get well over one millisecond -- one hundred times slower. This is why mainframes are still king for many orgs: They're the ultimate solution for dealing with speed-of-light delays. PS: I've seen multiple attempts to convert mainframe solutions to modern "racks of boxes" and it was hilarious to watch the architects be totally mystified as to why everything was running like slow treacle when on paper the total compute throughput was an order of magnitude higher than the original mainframe had. They neglected latency in their performance modelling, that's why!
  - le-mark 2y ago

rbanffy 2y ago

Maybe, but then you need to engineer the 99.99999% uptime yourself. If it were actually cheaper, IBM wouldn’t be selling these machines so well.

oldpersonintx 2y ago
[dead]
wolf550e 2y ago
I don't think they are actually selling those machines so well. They have a captive legacy customer base, who else is buying those?
- neverartful 2y ago
  From TFA: "Overall, Z is growing very healthily. LinuxONE is the fastest area of growth for us right now." However, he didn't elaborate or give any examples. If I were the interviewer, I would have followed it with: "Oh?! Can you provide some examples for the readers who believe that you only sell to captive audiences?"
  - rbanffy 2y ago
    If I had the need to ridiculously scale up a Linux workload, I’d immediately think LinuxONE or POWER. If I had the R&D budget for an on-prem VM hosting box, I’d seriously consider their smaller Express 4, which is not much pricier than a similarly capable x86 machine.
  - IanCutress 2y ago
    Because I've covered that in a previous video :) https://www.youtube.com/watch?v=ZVO55oxJ464 https://www.youtube.com/watch?v=ZVO55oxJ464
Muromec 2y ago
They are mostly selling to the captive audience who is 40 years deep into COBOL and can't pull out until it falls on top of them.
- feurio 2y ago
  Well, you know what to do - just sell a system that runs COBOL on an x86 cluster or Kubernetes or whatever, and you'll make billions! Be sure to allocate me a bunch of shares for giving you the idea.
- pjmlp 2y ago
  Many of those machines are running Java workloads, also COBOL isn't the only mainframe language, and the best thing in terms of security is that they don't use C as systems language, rather saner stuff with proper arrays, strings and bounds checking. That is why Unisys ClearPath MCP is still a thing, tracing back to its Burroughs 1961 heritage, security above all.
  - rbanffy 2y ago
    Java workloads might be easier to port away from but aren’t MCP machines x86-based nowadays? Unisys did a ton of work for Microsoft making itself obsolete by adding a lot to their secret multiprocessor stuff to Windows. It’s amazing how Microsoft convinced so many companies to shoot their feet.
    - pjmlp 2y ago
      They still use NEWP as systems language, there is no C stuff to exploit.
      - rbanffy 2y ago
        I had a lot of trouble with the idiosyncratic ways of MCP, and I have no doubt Tron's villain was named after it.
- rbanffy 2y ago

neverartful 2y ago

These comments always come up with every mainframe post. It's not only about performance. If it were it would be x86 or pSystems (AIX/POWER). The reason customers buy mainframes is RAS (reliabililty, availability, scalability). Notice that performance is not part of RAS.

jiggawatts 2y ago
You and the parent are both "missing the point", which is sadly not talked about by the manufacturer either (IBM). I used to work for Citrix, which is "software that turns Windows into a mainframe OS". Basically, you get remote thin terminals the same as you would with an IBM mainframe, but instead of showing you green text you get a Windows desktop. Citrix used to sell this as a "cost saving" solution that inevitably would cost 2-3x the same as traditional desktops. The real benefit for both IBM mainframes and Citrix is: latency. You can't avoid the speed of light, but centralising data and compute into "one box" or as close as you can get it (one rack, one data centre, etc...) provides enormous benefits to most kinds of applications. If you have some complex business workflow that needs to talk to dozens of tables in multiple logical databases, then having all of that unfold in a single mainframe will be faster than if it has to bounce around a network in a "modern" architecture. In real enterprise environments (i.e.: not a FAANG) any traffic that has to traverse between servers will typically use 10 Gbps NICs at best (not 100 Gbps!), have no topology optimisation of any kind, and flow through at a minimum one load balancer, one firewall, one router, and multiple switches. Within a mainframe you might have low double-digit microsecond latencies between processes or LPARs, across an enterprise network between services and independent servers its not unusual to get well over one millisecond -- one hundred times slower. This is why mainframes are still king for many orgs: They're the ultimate solution for dealing with speed-of-light delays. PS: I've seen multiple attempts to convert mainframe solutions to modern "racks of boxes" and it was hilarious to watch the architects be totally mystified as to why everything was running like slow treacle when on paper the total compute throughput was an order of magnitude higher than the original mainframe had. They neglected latency in their performance modelling, that's why!
- le-mark 2y ago