On Sat, Jun 25, 2011 at 11:57 AM, Gregory D'alesandre <[email protected]>wrote:
> Hi Nick, I've answered your questions below as best I can... > > On Sat, Jun 25, 2011 at 4:40 AM, Nickolas Daskalou <[email protected]>wrote: > >> Hi Greg, >> >> Thanks for putting together the update. I have some questions, if you >> wouldn't mind answering them: >> >> >> *(1) Why no RAM-charge?* >> >> The argument that a CPU-charge was not reflective of an application's real >> resource usage is understandable, and that the underlying reason for >> changing the model was because RAM usage was not taken into account with the >> old model. Why not then just simply *add a RAM-charge* to account for >> this, instead of flipping the ecosystem on its head and drastically changing >> the pricing model, as you have now done? This question has been asked many >> times since the new pricing model was released and it is yet to receive an >> official answer. >> > > We have in essence added a RAM charge by charging for Instances. By having > an Instance up with an allocated amount of memory you are essentially using > that RAM. So, by charging for the Instance we are charging you for the > combination of the RAM and CPU. We considered splitting this charge out so > we would continue to charge CPU-hours and then also charge Instance-hours > (which we could've called RAM-hours). This both seemed more confusing as > well as would not have been cheaper, so it didn't seem worthwhile. I know > that there has been a lot of discussion about charging this way instead. In > the end it, whether you call it RAM-hours or Instance-hours and whether or > not you charge for CPU-hours on top of it, it would end up with the same > result. Which is that applications are charged for the amount of RAM > allocated and the amount of time it is allocated for. This means > applications that want to save money will need to optimize around using > fewer RAM-hours which in essence means taking less time to get things done. > But I might be misunderstanding the question because if you feel a RAM > charge is straightforward I'm not sure I understand why you feel charging > Instance-hours is "flipping the ecosystem on its head." I hope that helps > give some clarity. > I guess the differences are as follows: 1. With instance hours the focus is not on optimizing RAM consumption at all, but on reducing latency (increasing RAM consumption, reducing costs) and controlling when instances come up and go. 2. With RAM-hours pricing, idle instances would cost lesser as CPU-hours are not charged. 3. The pricing would be linear and pay-for-what-you-use: amount of RAM for the time used. Not a fixed instance type with some amount of RAM. So, it is "flipping the ecosystem on its head". -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.
