To lock-in or not to lock-in

Controversy talk about vendor lock-in and it implications.

Transcript

1. www.softserve.ua To lock-in or not to lock-in Copyright © 2013 SoftServe, Inc. by Sasnyk Petro @stoune stoune
2. www.softserve.ua • The promise of cloud computing • “Magic” numbers • Alternatives • How to mix best of two worlds • Conclusions Contents
3. www.softserve.ua 1. No upfront costs 2. Economic of scale 3. Flexibility 4. Infinite Elasticity 5. SLA 6. Automation The promise of cloud computing
4. www.softserve.ua 1. Relatively slow 2. Expensive 3. Lots of hidden costs 4. Have a tons of useful, but proprietary features The promise of cloud computing: AWS in reality
5. www.softserve.ua “Magic” numbers
6. www.softserve.ua 1. Digital Ocean 2. ProfitBricks 3. GoGrid 4. Linode 5. …. Meet new generation of IaaS providers
7. www.softserve.ua Digital Ocean vs AWS: 10x Performance For 1/3 Cost link
8. www.softserve.ua Generic solutions • Easy to migrate • More expensive development • Cheaper operation • Require in-house expertise Vendor specific solutions • Faster time-to market • Expensive operation • Tricky terms and conditions Vendor specific services vs generic solutions
9. www.softserve.ua 1. Listen your customer and select service according to it’s current and future needs, but not for the case that you suddenly became “new Facebook” in one night. 2. Be realistic 3. Think, measure then deploy. Don’t follow blindly the crowd, consider different alternatives. 4. Use TCO calculators (http://serverbear.com/ etc.) Conclusions
10. www.softserve.ua 5. Do not lock-in without need (First pill for free) 6. Consider alternatives 7. Use opensource tools which you recommend for itself: 1. Vagrant 2. Chef/Puppet/Ansible/… 3. OpenStack 4. MySQL/PostgreSQL/MongoDB/Cassandra 5. Nagios/Zabbix Conclusions
11. www.softserve.ua