5 ms·

It reminds me of https://www.frankmcsherry.org/graph/scalability/cost/2015/01/15/COST.html https://www.frankmcsherry.org/graph/scalability/cost/2015/01... - you

It reminds me of https://www.frankmcsherry.org/graph/scalability/cost/2015/01/15/COST.html https://www.frankmcsherry.org/graph/scalability/cost/2015/01... - you pay an extremely high cost upfront without even noticing - for the promise that at some point, you'll be able to scale out to any scale, with no extra human effort and for costs proportional to the scale. I have not observed enough uses of serverless I can draw conclusions from, but if it's anywhere like hadoop style scaling, 95% of the users will pay 10x-100x on every metric without ever actually deriving any benefit compared to a non-distributed reasonably defined system, and 5% will actually benefit -- but everyone would want to put it on their resume and buzzword-bingo cards.

m_mueller 6y ago
It depends. Give you an example: We have an expensive reporting workflow that needs considerable resources, but usually only once a month. Two ways to do this cost effectively: Scale down containers to zero, meaning several minutes of response time if you do need it ad-hoc sometime. At least as complex to configure. OR you fit it to AWS SAM. The latter has proven to be a good match for us. I solved the debugging story by keeping lambda functions AWS agnostic and wrapping them i a Flask debug server - will respond exactly the same but it is all local in a single process.
- rytemporal 6y ago
  I’m the head of product for Temporal. You should check us out, have to self host (MIT) right now but we are working on a cloud. You can safely write code that blocks for years. No more job queues.
  - beagle3 6y ago
    I suppose you mean temporal.io, but just FYI, the first google result for me was https://github.com/RTradeLtd/Temporal https://github.com/RTradeLtd/Temporal which is far but not far enough to be “obviously unrelated”. You might want to give a link next time to avoid ambiguity.

m_mueller 6y ago

It depends. Give you an example: We have an expensive reporting workflow that needs considerable resources, but usually only once a month. Two ways to do this cost effectively: Scale down containers to zero, meaning several minutes of response time if you do need it ad-hoc sometime. At least as complex to configure. OR you fit it to AWS SAM. The latter has proven to be a good match for us. I solved the debugging story by keeping lambda functions AWS agnostic and wrapping them i a Flask debug server - will respond exactly the same but it is all local in a single process.

rytemporal 6y ago
I’m the head of product for Temporal. You should check us out, have to self host (MIT) right now but we are working on a cloud. You can safely write code that blocks for years. No more job queues.
- beagle3 6y ago
  I suppose you mean temporal.io, but just FYI, the first google result for me was https://github.com/RTradeLtd/Temporal https://github.com/RTradeLtd/Temporal which is far but not far enough to be “obviously unrelated”. You might want to give a link next time to avoid ambiguity.