2d:null 2e:["$","title",null,{"children":"Serverless | Jonas Almeida | Observable"}] 33:Tc9b,#### The I can't be bothered Execution model If the code places no explicit requirements on the operating system, then a Function-as-a-Service is avaliable in all of the major cloud computing companies, from [Google](https://cloud.google.com/functions), [AWS](https://aws.amazon.com/lambda) and [Microsoft](https://azure.microsoft.com/en-us/services/functions/#overview) to edgier [Cloudflare](https://cloudflareworkers.com) [[dashboard](https://dash.cloudflare.com)] and [Akamai](https://www.akamai.com/products/serverless-computing-edgeworkers). The resulting managed service solution is observed to be of up to **two orders of magnitude**. #### The no server service The modern web server supports natively I/O operations like range-requests that allow for Web APIs to be entirely developed in the client side, as exemplified here for [digital pathology](https://episphere.github.io/imagebox3) and [genomics](https://episphere.github.io/vcf). Note two years ago a general purpose portable [WASM sandbox](https://www.w3.org/2019/12/pressrelease-wasm-rec.html.en) became available to help. #### Data, data everywhere, and not a drop to drink? We don't control other people's development of APIs, this causes them to converge ... here's an example: RDFtable, and this is before even considering those that can be assembled on the client side if the serialization allows it. #### Serialization before metadata Data formats without a traversable (supporting range-requests) index are an unnecessary logistic obstacle to serverless approaches. This is true for applications in Bioinformatics as it is for Physics and other fields generating Data Commons / Big Data. Examples: Apache's Parquet, Google protocol buffers, HDF5, Zaar ... #### When only 4% are reproducible and it is not safe Observe and collaborate - https://www.nature.com/articles/d41586-021-01174-w, maybe start with Observable. #### AuthZ parsing FAIR as an R that starts with a A: client-side authentication orchestration with server-side scopping authorization. (recall RAS). #### SDK - the convenience of BYOC Cloud services like [BigQuery provide SDKs](https://cloud.google.com/bigquery/docs/reference/libraries) (*software development kits*, i.e. libraries/toolboxes) that operate the API of the remote environment from the convenience of your favorite analytical environment. See, for example, https://solutions.rstudio.com/db/databases/big-query --> https://bigrquery.r-dbi.org. These SDKs include negotiating authentication and authorization, enabling a scalable Bring Your Own Client (BYOC) approach to the design and management of the data infrastructure. If you are an analyst who wants to understand one single topic from this list, this will be the most relevant. 34:T69e,## Examples for Discussion url [openHealth] url [preprint] url url url url url OpenHealth url,twitter https://mathbiol.github.io/dsripMap/#ACSNY2015.csv 35:[] 31:{"id":"0f87895a1159da7c","slug":"serverless","title":"Serverless","thumbnail":"eddfe74d9e35b534b665d36c0d984e1ac7ff14b058016cd707ff44725cba0898","access_level":"public","create_time":"$D2022-07-12T12:49:59.390Z","update_time":"$D2022-10-14T14:54:15.385Z","view_time":null,"first_public_time":"$D2022-07-12T12:50:21.240Z","fork":false,"stars":1,"starred":false,"owner":{"type":"team","id":"0f2e596d5bd6ef6e","login":"episphere","name":"Jonas Almeida","avatar_url":"https://avatars-next.observableusercontent.com/3a0ff5abff6da0cdbd100da305ba323e12d43c8fcb7604275e6c72591db9dc69","tier":"starter_2024"},"version":97,"publish_version":null,"role":null,"authored":false,"authors":[{"type":"individual","id":"4355fe13cb71b11d","login":"episphere","name":"Jonas Almeida","avatar_url":"https://avatars-next.observableusercontent.com/3a0ff5abff6da0cdbd100da305ba323e12d43c8fcb7604275e6c72591db9dc69","tier":null}],"members":[{"type":"individual","id":"4355fe13cb71b11d","login":"episphere","name":"Jonas Almeida","avatar_url":"https://avatars-next.observableusercontent.com/3a0ff5abff6da0cdbd100da305ba323e12d43c8fcb7604275e6c72591db9dc69","tier":null,"role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":null,"type":"notebook","body":{"title":"Serverless","cells":[{"id":0,"value":"# Serverless\n\n

Jonas Almeida, Daniel Russ, Jeya Balasubramanian, Praphulla Bhawsar

\n\n\nNotes for discussion of *\"serverless APIs\"* at NIH [Office of Data Science Strategy](https://datascience.nih.gov/about/odss) technical implementation working group meeting on July 13th 2022.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":12,"value":"### What is serverless?\nIt is **a)** a [cloud execution model](https://en.wikipedia.org/wiki/Serverless_computing), and it is **b)** a [FAIR approach to Data Commons](https://jonasalmeida.github.io/preprints/Montserrat_Garci%CC%81a-Closas_2022_Moving_Towards_FAIR_preprint.pdf) (see also section *3. Discussion* in [arXiv:2207.01734](https://arxiv.org/pdf/2207.01734.pdf)). It is also the edge Cloud Computing is evolving to reach as distributed computing, where the code, rather than the data, does the traveling. Specifically, with interoperability and [AuthZ](https://episphere.github.io/oauth) (show equivalent for an NCI specific middle layer, [epibox](https://observablehq.com/@episphere/epibox)) resolved at the HTTP call level with **zero idle** on the server-side and **zero-footprint** on the client side.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":18,"value":"$33","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":40,"value":"$34","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"}],"files":"$W35","stdlib":"1","titling":"auto"}} 32:"$31:body"