#AmazonEFS

[govindhtech]

Prebid Server Deployment Now Available on AWS

Prebid Server Deployment on AWS is now generally available, according to a statement from Amazon Web Services (AWS). With this new solution, AWS customers can quickly and easily deploy Prebid Server, an open-source real-time ad monetization solution, in their own Amazon Private Cloud (VPC) environment, giving them complete control over the auction decision logic.

This allows customers to monetize their content at scale with programmatic advertising across web, mobile, connected TV, gaming, and other properties. For sell-side customers seeking more customisation than proprietary solutions can offer, Prebid Server Deployment on AWS streamlines “header bidding” workflows, which are content providers bidding for ad space to advertisers across multiple third-party advertising platforms simultaneously.

Many AWS customers use third-party ad platforms, where advertisers bid in real-time auctions for the chance to display their advertisement to consumers, to monetize content through advertising across web, mobile, connected TV, gaming, and other assets. These AWS clients use “header bidding” technology, which enables them to scale access to numerous advertisers’ bids and concurrently auction ad space across several ad platforms, in order to optimise income potential.

The customer experience may be impacted by the way that existing header bidding operations are implemented in the browser using JavaScript. This is addressed by server-to-server methods like the Transparent Ad Marketplace (TAM) offered by Amazon Publisher Services. It is a difficult task for even well-resourced organisations to deploy and manage architectures that are geographically distributed and can support high volumes of traffic with low latency, so publishers who want to customise their auctions or run them in their own AWS accounts are losing out on potential revenue.

A Quick Overview of Prebid Server Deployment on AWS Prebid Server Deployment on AWS, which is generally accessible today, lets users install Prebid Server in their own Amazon VPC environment in a matter of simple steps. This new AWS solution offers pay-as-you-go economics and global scalability along with complete control over auction decision logic and how clients stream transaction data, including necessary revisions.

For AWS users wishing to establish communication with ad platforms, the release of Prebid Server Deployment on AWS signifies a significant advancement as Prebid Server offers connectivity to more than 180 advertising platforms. Prebid Server has been used as a jumping off point by customers with mobile and CTV applications as well as commerce media firms to construct that link. Prebid Server Deployment on AWS speeds up the development process so that customers may have their own monetization engine sooner.

“AWS customers can own their monetization stack with Prebid Server Deployment,” stated Tim Barnes, General Manager of Advertising and Marketing Technology Solutions at AWS. Customers may maximise revenue by holding auctions across hundreds of partners by using this solution, and they can access all of the resulting data in their own AWS environment. They can now develop novel approaches to decision-making on monetization, and they are just need to cover the costs of the supporting infrastructure.

In order to provide AWS clients with ad-supported content properties with scalable ad income access through Prebid.org’s network of advertising platforms, Prebid Server Deployment on AWS was created as a click-to-deploy solution. AWS joined the open-source community that advances omni-channel ad tech auction standards in June by joining Prebid.org.

With the help of this new solution, AWS offers a thorough deployment template along with support for databases, caches, data pipelines, load balancing, containers, and Amazon Elastic Compute Cloud (EC2) instances running the most recent version of Amazon S3. Prebid Server Deployment on AWS can also handle up to 100,000 requests per second (RPS) on average, and it records transaction and operational logs from every Prebid Server container for reporting, analysis, and monitoring purposes. Delivering the operational quality, efficiency, and cost optimisation that AWS customers need, it is a secure, dependable, and sustainable solution for content monetization.

Under the Hood of Prebid Server Deployment on AWS With just one click in the AWS Management Console or AWS Solutions Library, Prebid Server Deployment on AWS offers the complete infrastructure needed to host a Prebid Server with production-grade availability. For admission into Prebid Server Deployment on AWS, Prebid.js code on a customer’s content property secretly transmits ad buyer bid requests to the web application firewall endpoint of Amazon CloudFront/WAF. The load balancer receives this request and forwards it to a cluster of Prebid servers that are operated by Amazon Elastic Container Service (Amazon ECS).

The inbound request is then verified and improved by Prebid Server, which also contacts server-side bid adapters in parallel. In the meantime, in order to monitor the overall health of the solution, Amazon CloudWatch gathers metrics from different resources that are engaged in handling requests and answers through the solution and alerts that are used to scale-in or out each container cluster. Each container’s Prebid Server runtime and metrics logs are momentarily kept on a shared Amazon Elastic File System (Amazon EFS).

Read more on Govindhtech.com

[awscloud4u]

#AmazonEFS introduces a new storage class that automatically reduces cost by up to 85% for infrequently accessed files! https://t.co/7cjIrjd5F4 https://t.co/yOy68cKitQ (via Twitter http://twitter.com/awscloud/status/1095789117810663429)

[awsexchage]

AWS Price List API から価格情報を引いてくるツールを作った http://ift.tt/2w0X2AE

AWS Price List API で雑に月額を出してみる の続きです。

y13i/awsprice: CLI for AWS Price List API.

今回はある程度汎用的に使えそうなコマンドラインツールの形式で、 Go で実装しました。

実装について

Price List API のオファーファイルは本当にデカいので毎回落としてくるのは厳しいと思い、キャッシュする機構をつけてみました。

macOS の場合は /var/folders/xx/xxxxxxxxxxxxxxxxxx/T/awsprice らへんに保存されると思います。

Homebrew でインストールできるようにもしてみました。

使用したライブラリなど

表示には k0kubun/pp (Goのデバッグに便利なpretty printライブラリ「pp」を作った) を使わせてもらいました。大変きれいにターミナルに出力されて感激です。

CLI の骨組みは urfave/cli を使わせてもらいました。これがなかったらオプションまわりとか面倒すぎて死ぬ気がしました。

スライスをソートするのが面倒だったので bradfitz/slice も使わせてもらいました。 Go 力が足りない自分にはありがたい…

使いみちについて

おおよその使い方は README.ja.md に書いたので、実際何に使えるか考えてみます。

EFS が東京にまだ来てない現状を確認してみる

$ awsprice listOfferRegions -o AmazonEFS []awsprice.RegionCode{ "ap-southeast-2", "eu-west-1", "us-east-1", "us-east-2", "us-west-2", }

まだかな〜

リージョン毎に使えるインスタンスタイプの違いを見てみる

$ awsprice -f json listAttributes -p "Compute Instance" -r us-east-1 | jq ".instanceType | length" 77 $ awsprice -f json listAttributes -p "Compute Instance" -r ap-northeast-1 | jq ".instanceType | length" 73 $ awsprice -f json listAttributes -p "Compute Instance" -r eu-central-1 | jq ".instanceType | length" 56 $ awsprice -f json listAttributes -p "Compute Instance" -r ap-south-1 | jq ".instanceType | length" 45

新造リージョンは古いインスタンスタイプがなかったりで、使えるインスタンスタイプが少ないことがわかります。

awsprice 単体はそれほどクエリー機能が強力ではないですが、 JSON で出力できるので jqなどとの連携が簡単です。

最強インスタンス3年一括お買い上げの値段で腰を抜かしてみる

$ awsprice -f json listProductTerms -r ap-northeast-1 -a instanceType=x1.32xlarge -a operatingSystem=Linux -a preInstalledSw=NA -a tenancy=Dedicated --termType Reserved -t "PurchaseOption=All Upfront" -t LeaseContractLength=3yr -t OfferingClass=standard | jq ".[0].OfferTerms[0].PriceDimensions" { "6BKDT7UNN25Z7CYZ.NQ3QZPMQV9.2TG2D8R56U": { "RateCode": "6BKDT7UNN25Z7CYZ.NQ3QZPMQV9.2TG2D8R56U", "Description": "Upfront Fee", "BeginRange": "", "EndRange": "", "Unit": "Quantity", "PricePerUnit": { "USD": "142204" }, "AppliesTo": [] }, "6BKDT7UNN25Z7CYZ.NQ3QZPMQV9.6YS6EN2CT7": { "RateCode": "6BKDT7UNN25Z7CYZ.NQ3QZPMQV9.6YS6EN2CT7", "Description": "USD 0.0 per Linux/UNIX (Amazon VPC), x1.32xlarge instance-hour (or partial hour)", "BeginRange": "0", "EndRange": "Inf", "Unit": "Hrs", "PricePerUnit": { "USD": "0.0000000000" }, "AppliesTo": [] } }

1,600万円あったら何するかなあ…

最強インスタンスでなくても、この要領でインスタンス単体のお値段を調べることができます。

で、月額は？

汎用的に使えるツールを目指した結果、 EC2 以外のサービスの情報も引けるようになったのはいいのですが、価格単位が時間以外のものも混ざってきて、しかも時間単位でも RI の扱いをどうしたものか……と悩んで結局実装してません。

なにかいいアイデアがあれば教えてください…

元記事はこちら

「AWS Price List API から価格情報を引いてくるツールを作った」

August 01, 2017 at 04:00PM