gcp-deployment-memcached-lite

Deployment automation of memcached-lite service on Google cloud VMs

Requirement

• Make sure GCP SDK is installed along with GCP Beta SDK (for linking projects with billing account)

Config File (config.yml)

• Contains initial values for network, firewall, server name, client name, zone and scripts path.

  Modify the config file as you require.

Operations performed in order

1. gcloud web authentication

2. Project creation/selection and then setting the current project to default configuration

3. API enabling – cloudresourcemanager and compute

4. VPC/Network creation

5. Firewall opening for tcp,udp,icmp and tcp:22,tcp:3389 ports

6. Two VM creation along with startup script (metadata.sh) for library installation

7. Waiting for startup scripts to finish on each VM

8. Server code Installation/Copying from local machine

9. Client code Installation/Copying from local machine

10. SSH into client and server VMs and unzipping code

11. Server startup

12. Client testing – performs 4 client tests

13. Copying server and client logs from VM to local machine

14. VM Shutdown/Stopping

Run Script

• create.sh

• cleanup.sh

  Resource cleanup – Deletes VM, Firewalls, and Network in that order

Cost

• Virtual Machines

  • n1-standard-1 (1 vCPU, 3.75 GB memory) => $0.04749975 per hour * 2

• Network

  • Ingress traffic => No charge for ingress traffic

  • Egress to the same Google Cloud zone when using the internal IP addresses of the resources1 => No charge

• Firewall

  • 500 or fewer attributes in the policy (standard) => $1 per VM covered by the policy

  There are two VMs which uses the firewall and hence it will cost $2 per month.

VM Cost Estimator for the selected VMs

Performance of key-value store when deployed to cloud VMs

All results can be found in the output-logs folder along with server logs after script execution

1. 100 simultaneous connections to server and closing them afterwards (CONNECTION TEST)

   Test 1 Execution Time: 61743.148051 milliseconds

   ---

2. 100 connections to server and each perform 1 write and then 1 read

   Test 2 Execution Time: 61758.51037199999 milliseconds

   ---

3. 1 client writes to 1 key every 0ms and another client reads from the same key every 0ms for 100 times (Performance + Concurrency test)

   Test 3 Execution Time: 2175.790368999995 milliseconds

   ---

4. 2 clients write to same key at the same time for 500 times (Concurrency test for DB and multiple requests from a single client test)

   Test 4 Execution Time: 4209.0177080000285 milliseconds