What is Severless?
Serverless computing is a way to provide back-end services only when it is in use. Serverless providers allow users to write and deploy code without worrying about the underlying infrastructure. Companies that use back-end services from serverless vendors will be billed according to usage. The service is auto-scaling, so you don’t have to reserve and pay for a fixed amount of bandwidth or number of servers. Despite the name serverless, physical servers will continue to be used, but developers will not need to be aware of them.
Difference from cloud computing
Cloud computing
Serverless
Traditional cloud computing allows you to remotely rent a fixed number of servers or a fixed amount of server space. Developers and businesses renting these fixed units of server space are generally overpurchased to prevent spikes in traffic and activity from exceeding monthly limits and corrupting applications. This means that much of the server space paid can be wasted. Cloud vendors have introduced an auto-scaling model to address this issue, but even with auto-scaling, spikes in unwanted activity, such as DDoS attacks, can be very expensive.
Serverless computing allows developers to purchase back-end services with a flexible “pay-as-you-go” approach. This means that developers only have to pay for the services they use. It’s like switching from a mobile phone data plan with a fixed monthly plan to a plan that charges only for data used.
Pros
- 市場投入までの時間の短縮とソフトウェアリリースの迅速化
- 運用および開発コストの削減
- スケーリングするためのより小さなコスト
- ソフトウェアの複雑さを軽減します
- パッケージングと展開を簡素化し、システム管理は必要ありません。
- メンテナンスコストゼロ
Cons
- 各ベンダーのサーバーレステクノロジーにより、コードはそのシステムに適合するように設計する必要があります。
- コールドスタートは、サーバーレスインフラストラクチャが最初のリクエストを満たすために内部リソースを初期化する必要がある場合の問題です
- プロバイダーは、同じ物理サーバー上で異なる顧客からのコードを実行する場合があります（これはマルチテナンシーとも呼ばれます）
Serverless is a new approach for creating and deploying applications that allow developers to focus on writing code. Reduce time-to-market, operational costs, and system complexity. Eliminates the need to set up and configure physical servers or virtual machines by leveraging third-party services such as AWS Lambda, but locks applications and their architecture to specific service providers.