3 ms·

The question should be more why would there have to be such large batches which can be SPOF for so many doses. Wouldn't it make more sense to make smaller seper

The question should be more why would there have to be such large batches which can be SPOF for so many doses. Wouldn't it make more sense to make smaller seperate batches to derisk the process. This is akin to me using a large 64 core server for an app, instead of say 4x16 . Sure, the larger server can easily handle the load, however the 4x16 setup is more redundant and avoids many scenarios where all users will be out of service Just because there are high capacity machines capable of doing that much does not mean it should be used especially a With a new process trying to scale up

tomaskafka 6y ago
You might have smaller batches and smaller throughput, but some pesky people keep shouting that we need all the vaccines yesterday :).
maxerickson 6y ago
The production goal is hundreds of millions of doses this year, so I guess I wouldn't want to speculate too much about what batch size is sensible.
- manquer 6y ago
  eventual goal maybe, Not for the month of March. 15M out of planned 20M ? . The scale out for any process starts small to try avoid this kind of issues. Yes fuck ups happen at plant and in planning, and yes we are not experts in the field to comment in detail on what is best approach. However if you are looking at it as an admin, clearly risk was not well managed.
fragmede 6y ago
To use the same alarmism and a software analogy: sw developer destroys thousands of hours of CPU time with bad code!1 Yet every single time I push, the CI pipeline runs an extensive suite of automated tests, and I'd not have it any other way. Wait but CPUs aren't being destroyed by running tests, you say. Yeah but niche reporting is always atrocious and a mainstream writer can't be trusted to get things entirely right. (If you're using a 4x16 server, whatever that is, to provide isolation in your system, you may want to talk to additional devops folks for other options.)

tomaskafka 6y ago

You might have smaller batches and smaller throughput, but some pesky people keep shouting that we need all the vaccines yesterday :).

maxerickson 6y ago

The production goal is hundreds of millions of doses this year, so I guess I wouldn't want to speculate too much about what batch size is sensible.

manquer 6y ago
eventual goal maybe, Not for the month of March. 15M out of planned 20M ? . The scale out for any process starts small to try avoid this kind of issues. Yes fuck ups happen at plant and in planning, and yes we are not experts in the field to comment in detail on what is best approach. However if you are looking at it as an admin, clearly risk was not well managed.

fragmede 6y ago

To use the same alarmism and a software analogy: sw developer destroys thousands of hours of CPU time with bad code!1 Yet every single time I push, the CI pipeline runs an extensive suite of automated tests, and I'd not have it any other way. Wait but CPUs aren't being destroyed by running tests, you say. Yeah but niche reporting is always atrocious and a mainstream writer can't be trusted to get things entirely right. (If you're using a 4x16 server, whatever that is, to provide isolation in your system, you may want to talk to additional devops folks for other options.)