Backend Architecture and Component Selection for Startup Companies

Introduction – The article discusses the importance of building a backend architecture for startup companies and defines the four layers of a backend stack: programming languages, components, processes, and systems.

Backend Stack Overview – A diagram (Fig.1) illustrates the stack, and Fig.2 shows its structure.

Component Selection – Detailed recommendations are given for each category:

Project/Bug management: Redmine, Phabricator, Jira, Wukong CRM.

DNS: Alibaba Wanwang, Tencent DNSPod, Amazon Route 53.

Load Balancer: Cloud provider LB (SLB, CLB, ELB) or LVS+Nginx.

CDN: Domestic providers (Wangsu, Tencent, Alibaba) and overseas (Amazon CloudFront, Akamai).

RPC frameworks: cross‑language (Thrift, gRPC, Hessian, Hprose) and service‑governance (Dubbo, DubboX, Motan, rpcx).

Service discovery: etcd, Consul, ZooKeeper.

Databases: Traditional (MySQL/MariaDB, Oracle, PostgreSQL) and NewSQL (TiDB, CockroachDB).

NoSQL types: key‑value, column‑family, document, graph.

Message middleware: Kafka, RocketMQ, RabbitMQ, etc.

Logging: ELK stack (Elasticsearch, Logstash, Kibana, Filebeat).

Monitoring: Prometheus + Grafana, Zabbix, Open‑Falcon.

Configuration management: ZooKeeper/etcd or push‑based tools (Puppet, Ansible).

Release & deployment: Jenkins, GitLab CI, Walle, etc.

Jump server: Jumpserver.

Machine management: Ansible, SaltStack, Puppet, Chef.

Startup‑Specific Advice – Choose languages familiar to the team, prefer mature open‑source components, rely on reliable cloud services, establish clear development, release, and operations processes, and automate where possible.

Cloud‑Based Architecture – Combining the selected components on a cloud provider yields the final architecture diagram (Fig.10).

References – A list of URLs to documentation and articles is provided at the end of the original text.