这个AWS RI/SP模拟引擎是否有趣/有价值？
简而言之：我们开发的这个AWS RI/SP工具可能具有独特的功能，尤其对那些运行大型、高变动工作负载并且有高覆盖目标的公司来说非常重要。我们希望了解这一点的真实性，以及是否应该将其开放给其他人，甚至可能免费或作为开源软件（OSS）。
设置：我们是一家小型云成本咨询公司，而不是工具供应商。我们确实为自己的使用构建了许多内部工具。我们的客户是年支出在七位到九位数（美元）的公司。
在过去几年中，我们构建的一个内部工具是一个用于RI/SP的模拟引擎，可以执行多项功能，其中一些我们认为在当前市场上可能是独一无二的，并且对其他人可能有价值。
1.) 可视化任何承诺类型（任何RI、任何SP、任何形式）的整个承诺节省/折扣曲线，显示在每个承诺级别上实现的确切节省和边际折扣率。这解锁了我们认为的最佳购买策略：“购买SP美元或RI单位，直到下一个便士/实例小时的折扣低于10%”，而不是采取全有或全无的建议，使用平均折扣（例如，“通过以5%的整体折扣购买每小时10美元的SP来覆盖所有内容”）。
2.) 这样的节省/折扣曲线也可以叠加，以显示折扣率的变化如何解锁更多的覆盖空间。例如，决定用预付款购买部分或全部承诺会改变折扣率，这不仅是一个现金投资回报率的决策，还会改变曲线的形状，使得在相同风险水平下能够更有利可图地覆盖更多基础设施。
3.) 构建“假设”场景，叠加即将进行的计划购买，增加或减少覆盖使用，并过期不需要的承诺，以查看在各种假设下节省的变化。基本上，我们根据所需的回顾期的历史使用情况进行修改，以便进行前瞻性预测，并应用我们想要的任何当前或计划承诺。如有需要，我们可以同时运行多个这样的场景，以识别风险。
4.) 查看通过购买在现有SP之前应用的承诺如何提高节省。如果您的AWS控制台或首选工具当前仅显示有计算SP购买可能性，而RI和EC2实例SP的可能性很低或没有，您肯定处于这种情况。AWS并未显示当前占用覆盖空间的SP折扣如何可以沿着折扣曲线推移，这将解锁特定目标RI或EC2实例SP的购买，这些将首先应用并享有更高的折扣水平。
5.) 可以通过大语言模型（LLM）来驱动，以缩小在没有模拟可能性空间的情况下不可见的策略。重要的是，生成的数据是紧凑、易读且易于总结的。
在评论中添加几张带说明的图片，以帮助传达上述内容。
问题：
1.) 这些功能真的独特吗？还是有某些现有工具（无论是付费还是开源）基本上在做同样的事情？
2.) （如果第1个问题的答案是“是，独特”）这些功能是否足够有价值，以至于值得投入精力将其引入您的实践，还是它们仅仅是边际有趣（或者，可能，完全无趣）？
3.) 您认为模拟引擎可以用来解决其他相关问题吗？这些问题我在这里没有特别提到，但会更有价值？（我们有很多想法……策略回测、并购合并账单账户场景等。）
查看原文
TLDR; This AWS RI/SP tool we built might have unique features that are important especially for companies running large, high-variability workloads with high coverage targets. We want feedback on how true that is and whether we should open this up to others, possibly even for free or as OSS.<p>Setup: We're a small cloud cost consultancy, but not a tools vendor. We do build a lot of internal tooling for our own use. We serve companies spending 7- to 9-figures (USD) annually.<p>One of the internal tools we built over the past several years is a simulation engine for RIs/SPs that can do a number of things, some of which we believe might be one-of-a-kind in the space today and could be valuable to others.<p>1.) Visualize the entire commitment savings/discount curve for any commitment type (any RI, any SP, of any flavor), showing the exact savings achieved and the marginal discount rate at every level of commitment. This unlocks what we feel is the best purchasing strategy: "buy SP dollars or RI units until the next penny/instance-hour yields less than a 10% discount," rather than taking the all-or-nothing recommendation with an averaged discount (e.g., "Cover everything by buying a $10/hr SP at a 5% overall discount").<p>2.) Such savings/discount curves can also be stacked to show how changes in discount rate unlock more coverage space. For example, deciding to purchase some or all of a commitment with upfront dollars changes the discount rate, which is not just an ROI-on-cash decision, but also changes the shape of the curve to allow more infrastructure to be covered profitably at the same risk level.<p>3.) Build "what-if" scenarios that stack upcoming planned purchases, add or remove covered usage, and expire unwanted commitments to see how savings change under various assumptions. Essentially, we take the historical usage for the desired look-back period, modify it according to our whims for the forward projection, and apply any of the current or planned commitments that we want. We run multiple such scenarios simultaneously if needed to suss out risks.<p>4.) See how savings can be improved by purchasing commitments that would be applied before your existing SPs. If your AWS console or tool-of-choice currently shows you having only Compute SP purchase possibilities, and low or no RI and EC2 Instance SP possibilities, you're definitely in that boat. AWS doesn't show you how SP discounts that are currently eating the coverage space could instead be pushed out along the discount curve, which would unlock purchasing specific targeted RIs or EC2 Instance SPs that would get applied first and at higher discount levels.<p>5.) It can be driven by LLMs to narrow in on strategies that aren't visible without simulating the possibility space. Importantly, the data produced are compact, legible, and easily summarized.<p>Adding a couple of images in comment with explainers to help communicate some of the above.<p>Questions:<p>1.) Are these features truly unique? Or is there some existing tool (whether paid or OSS) essentially doing the same thing?<p>2.) (If answer to #1 is "yes, unique") Are these features enough of a value-add that they would be worth the effort to bring into your own practice, or are they only marginally interesting (or, perhaps, completely uninteresting)?<p>3.) Do you think the simulation engine could be used to solve other related problems that I'm not specifically suggesting here and that would be more valuable? (We have a lot of ideas... strategy back-testing, M&A merged billing account scenarios, etc.)