Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Copy of AWS Wicked-Rugby.com

Case Study
by

Ricky Chaggar

on 9 March 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Copy of AWS Wicked-Rugby.com

2) Challenges wicked-rugby.com By Martin Carlson
Account Manager Wicked-Rugby.com 1) What we know about
Wicked Rugby Wicked-Rugby.com Agenda 1) What we know (Wicked-Rugby.com)
2) Challenges
3) Cloud Computing
4) Why Amazon Web Services
5) Proposed Solution
6) Summary 3) What is Cloud Computing? "Cloud computing is the use of computing resources (hardware and software) that are delivered as a service over a network (typically the Internet)" "Cloud computing is a type of computing that relies on sharing computing resources rather than having local servers or personal devices to handle applications." Primary Cloud Computing Flavors
Premier Social Networking site for Rugby fans around the world Can create personal profile pages
Can create friend networks
Provides up to date News & Stats from leagues around the world User Base 50% Europe / 30% US / 20% Asia
2 Million unique users
Projected linear growth: 10 million by Feb 2013 wicked-rugby.com Current HW Infrastructure All servers reside in one location- Dallas, Texas 4 Red Hat Linux Servers (2 master: 1 active/1 passive. 2 slave)
-14GB RAM/ 2 dual core procs each
-160GB volume. 40GB replicated over each server Database cluster: Web/App cluster: 16 Red Hat Linux Servers. PHP/Apache in an open source MVC framework
-1.5GB RAM/ 1GHz Proc each F5 Load Balancing Hardware (refresh/update)(need comparable services) Akamai- Content Delivery Provider. For video and static content. (Contract expires 6 months) Damage Total: 25k/month
15k for hosting
10k Akamai a) Improve AVAILABILITY
b) Improve PERFORMANCE
c) ...All While Keeping COSTS DOWN 2M users 10M users Feb 2013 Feb 2014 12 months Cloud Computing Advantages Elastic Capacity Increase Business Speed & Agility Trade in Capital Expense for Variable Expense Fixed Capacity Demand 10M 2M (Users) 1 Month 12 Month Standard Capacity Model (Waste) (Waste) 2M 10M (Users) Month 1 Month 12 (Time) Elastic Capacity Demand Elastic Capacity Model Replace Costly IT budget with Utility model
Pay as you go, only for what you use!
PREVENT OVERSPENDING
No CapEx
No Longterm Commitments
No Upfront Costs
No HW costs
No Refresh Costs
No HW maintenance costs
-Dont Guess Capacity- How much time does it take you to go to market with an idea? How much money is wasted if experiment fails?.... Who takes the blame? Twice the Experiments= Twice the Success Rate Experiment Freely Go to Market Faster 4) A Cloud Computing Pioneer "We did a full evaluation of Amazon Web Services and a few other vendors and we found Amazon was in fact 3-4 years ahead of anyone else we had evaluated" -Lex Crossett, VP of Product Developement & CIO ...and we are still growing Pace of Innovation AWS Global Infrastructure 9 Regions
25 Availability Zones
38 Edge Locations
... and growing US - 30%
3 Regions
11 AZ's
18 Edge Locations Europe- 50%
1 Region (Ireland)
3 AZ's
12 Edge Locations Asia- 20%
3 Regions
7 AZ's
8 Edge Locations Security http://aws.amazon.com/security/ White papers: Security Processes, Risk and Compliance
Many Security Certifications: ISO 27001, PCI-DSS-1, etc
Customer penetration testing
Data privacy though encryption Our Datacenters are probably more secure than yours "Trusted by over 100,000+ customers, from the US Government to the largest global businesses in the world" Low Margin Philosophy Reduce Price More Customers More AWS Usage More Economy of Scale We need more infrastructure Volume Discounts Lower
Infrastructure
Costs 23 Price Reductions
Since 2006 December 2012 Elastic Cloud Compute EC2 AWS Core Components EC2 Amazons Virtual Machines (Instances)
On-demand compute power
Quickly Scale up or down
Easy Setup- Amazon Management Console
Select Region and AZ
Select OS, RAM, CPU, 32bit vs 64bit platform
qty Instances, name it, assign key pair, security group.... Launch away Servers start at $.02 (2 cents) an hour Auto Scaling -AWS Free service that allows you automatically scale up and down according to conditions that you define. Conditions can be based off of traffic, schedule Why is this important for you? Free service that prevents overspending, wasted capacity, unsatisfied customers. AWS Core Components Elastic Load Balancing EC2 #1 EC2 #1 EC2 #2 EC2 #1 -Automatically distributes incoming traffic across multiple EC2 instances to achieve greater availability. Can be done across multi-AZ's if need be. Unhappy Rugby Fans 5) Proposed Solution "With Amazon CloudFront, Earth Networks saved over 50% in CDN costs and anticipates 20% reduction in Amazon EC2 origin instances, without any reduction in performance compared to their previous CDN provider and without the need to write any customized code." “Scaling, provisioning, and security are very important for us and for our customers. AWS is a natural fit, because it allows us to have virtually unlimited capacity while only paying for what we use without upfront investments. AWS also makes it easy to deal with traffic spikes, which are common in sports, thanks to its elastic capabilities.” -Ramon Salvado, CTO of Gnuine (their IT provider) S3 Simple Storage Service RDS Relational Database Store -Prepackaged and managed DB instance
-MySQL, SQL server, Oracle
-Completely scalable
-Can set IOPS from 1,000-10,000
-Can set scheduled backup snapshots of DB and logs
-Can set multi-AZ's for increased availability -Object Based storage
-Highly Scalable
-1byte-5TB per object
-Keeps redundant copies across multi-AZ's for high availability
-easily accessible via URL
-highly secure via permissions, encryption, Access Control Lists etc. CloudFront -AWS's Content Delivery Service
-Integrated with AWS to provide low latency and fast transfer speeds to our many Edge locations
-Very Cost Effective
-Transfers Static & Dynamic content at the same low rate starting at $0.12/GB Route 53 -DNS Web Service
-Uses our Any Cast Network of DNS servers to automatically route your users to optimal locations to reduce latency.
-Designed to handle large query volumes
-Easy to manage your domains
-highly available Preliminary Web Architecture Wicked-Rugby.com Elastic Load Balancer EC2 Web Servers EC2 Web Servers EC2 Load Balancer EC2 Load Balancer Auto Scaling Group Auto Scaling Group Auto Scaling Group Auto Scaling Group RDS Master RDS Standby Custom Origin
Server (Dynamic) EC2 App Server EC2 App Server Custom Origin Server (Dynamic) Route
53 CloudFront S3 Availability Zone #1 Availability Zone #2 $15,000 /mo Hosting (co-location)
$10,000 /mo Akamai CDN (8TB~$1.25/GB) $28,282 /yr AWS 2 Million User Current Solution $25,000 /mo = $300,000 yearly
@ 2 Million Users (recommended web architecture) $11,520 /yr CloudFront $39,802 /yr $39,802 yearly
@ 2 Million Users $32,694 /mo Servers and storage (TCO calculator) $25,000 /mo $260,198 LESS Current Solution
Projections 10 Million Users -80 Web Servers
-16 DB servers
-5 TB of local storage
-Extended Akamai contract
-F5 Load Balancer Refresh
*No Availability *Hefty long term CDN contract *Bloated Data Center *Increased hosted prices *Susceptible to Performance issues $664,938 yearly
@ 10 Million Users $128,956 /yr AWS + $38,400 /yr CloudFront $167,356 yearly @ 10 Millions users $167,356 vs $664,938 a) Improve PERFORMANCE b) Improve AVAILABILITY c) Keep Costs Down Route 53 CloudFront Auto Scaling Load Balancing "PBS improved availability to over 99.99%" 6) Summary Next Steps -Separate Apache, PHP and MySQL 3-Tier vs All-in-one Single point of failure
Shared physical resources (RAM, CPU, HD) translates to waste.

Lets dive deeper into the technical side of our services
Lets challenge my TCO estimate
Lets determine a final configuration/architecture
Lets walk through the migration process
Thank you
Full transcript