Deep Dive Amazon EC2

1. © 2016, Amazon Web Services, Inc. or its Affiliates. All rights reserved.© 2015, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Steven Bryen - Manager,Solutions Architecture – Amazon Web Services
Chris Turvil - Head of Cloud and Platform Agility - Trainline
July 2016
Deep Dive on Amazon EC2 Instances

2. InstancesAPI
Networking
EC2
EC2
Purchase options
Amazon Elastic Compute Cloud is Big

3. Host Server
Hypervisor
Guest 1 Guest 2 Guest n
Amazon EC2 Instances

4. 2006 2008 2010 2012 2014 2016
m1.small
m1.large
m1.xlarge
c1.medium
c1.xlarge
m2.xlarge
m2.4xlarge
m2.2xlarge
cc1.4xlarge
t1.micro
cg1.4xlarge
cc2.8xlarge
m1.medium
hi1.4xlarge
m3.xlarge
m3.2xlarge
hs1.8xlarge
cr1.8xlarge
c3.large
c3.xlarge
c3.2xlarge
c3.4xlarge
c3.8xlarge
g2.2xlarge
i2.xlarge
i2.2xlarge
i2.4xlarge
i2.4xlarge
m3.medium
m3.large
r3.large
r3.xlarge
r3.2xlarge
r3.4xlarge
r3.8xlarge
t2.micro
t2.small
t2.med
c4.large
c4.xlarge
c4.2xlarge
c4.4xlarge
c4.8xlarge
d2.xlarge
d2.2xlarge
d2.4xlarge
d2.8xlarge
g2.8xlarge
t2.large
m4.large
m4.xlarge
m4.2xlarge
m4.4xlarge
m4.10xlarge
Amazon EC2 Instances History

5. What to Expect from the Session
• Getting the best system performance from your EC2
Instances
• How Amazon EC2 instances deliver performance
while providing flexibility and agility
• How to make the most of advanced features for EC2
and companion services

6. Defining Performance

7. • Servers are hired to do jobs
• Performance is measured differently depending on the job
Hiring a Server
?

8. Performance Factors
Resource Performance factors Key indicators
CPU Sockets, number of cores, clock
frequency, bursting capability
CPU utilization, run queue length
Memory Memory capacity Free memory, anonymous paging,
thread swapping
Network
interface
Max bandwidth, packet rate Receive throughput, transmit throughput
over max bandwidth
Disks Input / output operations per
second, throughput
Wait queue length, device utilization,
device errors

9. Resource Utilization
• For given performance, how efficiently are resources being used?
• Something at 100% utilization can’t accept any more work
• Low utilization can indicate more resources are being purchased
than needed

10. Example: Web Application
• MediaWiki installed on Apache with 140 pages of content
• Load increased in intervals over time

11. Example: Web Application
• Memory stats

12. Example: Web Application
• Disk stats

13. Example: Web Application
• Network stats

14. Example: Web Application
• CPU stats

15. • Picking an instance is tantamount to resource performance tuning
• Give back instances as easily as you can acquire new ones
• Find an ideal instance type and workload combination
Instance Selection = Performance Tuning

16. Delivering Compute Performance with
Amazon EC2 Instances

17. Review: C4 Instances
• Custom Intel E5-2666 v3 at 2.9 GHz
• P-state and C-state controls
Model vCPU Memory (GiB) EBS (Mbps)
c4.large 2 3.75 500
c4.xlarge 4 7.5 750
c4.2xlarge 8 15 1,000
c4.4xlarge 16 30 2,000
c4.8xlarge 36 60 4,000

18. Review: T2 Instances
• Lowest cost EC2 instance at $0.013 per hour
• Burstable performance
• Fixed allocation enforced with CPU credits
Model vCPU CPU Credits
/ Hour
Memory
(GiB)
Storage
t2.micro 1 6 1 EBS Only
t2.small 1 12 2 EBS Only
t2.medium 2 24 4 EBS Only
t2.large 2 36 8 EBS Only

19. How Credits Work
• A CPU credit provides the
performance of a full CPU core for
one minute
• An instance earns CPU credits at
a steady rate
• An instance consumes credits
when active
• Credits expire (leak) after 24 hours
Baseline rate
Credit
balance
Burst
rate

20. Tip: Monitor CPU credit balance

21. Delivering Memory Performance
with Amazon EC2 Instances

22. Announced: X1 Instances
• Largest memory instance with 2 TB of DRAM
• Quad socket, Intel E7 processors with 128 vCPUs
Model vCPU Memory (GiB) Local
Storage
x1.32xlarge 128 1952 2x 1920GB

23. Delivering I/O Performance with
Amazon EC2 Instances

24. Review: I2 Instances
16 vCPU: 3.2 TB SSD; 32 vCPU: 6.4 TB SSD
365K random read IOPS for 32 vCPU instance
Model vCPU Memory
(GiB)
Storage Read IOPS Write IOPS
i2.xlarge 4 30.5 1 x 800 SSD 35,000 35,000
i2.2xlarge 8 61 2 x 800 SSD 75,000 75,000
i2.4xlarge 16 122 4 x 800 SSD 175,000 155,000
i2.8xlarge 32 244 8 x 800 SSD 365,000 315,000

25. Device Pass Through: Enhanced Networking
• SR-IOV eliminates need for driver domain
• Physical network device exposes virtual function to
instance
• Requires a specialized driver, which means:
• Your instance OS needs to know about it
• EC2 needs to be told your instance can use it

26. Hardware
Before Enhanced Networking
Driver Domain Guest Domain Guest Domain
VMM
Frontend
driver
Frontend
driver
Backend
driver
Device
Driver
Physical
CPU
Physical
Memory
Network
Device
Virtual CPU
Virtual
Memory
CPU
Scheduling
Sockets
Application
1
23
4
5

27. Hardware
After Enhanced Networking
Driver Domain Guest Domain Guest Domain
VMM
NIC
Driver
Physical
CPU
Physical
Memory
SR-IOV Network
Device
Virtual CPU
Virtual
Memory
CPU
Scheduling
Sockets
Application
1
2
3
NIC
Driver

28. Tip: Use Enhanced Networking
• Highest packets-per-second
• Lowest variance in latency
• Instance OS must support it
• Look for SR-IOV property of instance or image

29. Leveraging Features of EC2 and
Companion Services

30. Auto Recovery for Amazon EC2
• Recover Instances that have
become impaired due to
underlying hardware problem.
• Instance maintains Instance ID,
Private IP, Elastic IP and
metadata.
• Configured through CloudWatch
EC2 Action

31. Auto Recovery for Amazon EC2
• Examples of problems causing system status checks to fail:
• Loss of Network Connectivity
• Loss of system power
• Software issues on the physical host
• Hardware issues on the physical host.
• Only supported on:
• C3, C4, M3, M4, R3, T2 and X1 Instances
• Instances in a VPC
• Instances with shared Tenancy
• Instances that use EBS storage exclusively

32. Auto Scaling – Lifecycle Hooks
• Hold instance in Pending or Terminating state.
• Notification of Lifecycle event triggering via
CloudWatch Events or SNS. (Lambda)
• Default timeout is one hour.
• Can CONTINUE or ABANDON. Set a default results
using –default-result

33. Auto Scaling Lifecycle Hooks - Adding
aws autoscaling put-lifecycle-hook
--lifecycle-hook-name my-hook
--auto-scaling-group-name my-asg
--lifecycle-transition autoscaling:EC2_INSTANCE_LAUNCHING
aws autoscaling put-lifecycle-hook
--lifecycle-hook-name my-hook
--auto-scaling-group-name my-asg
--lifecycle-transition autoscaling:EC2_INSTANCE_TERMINATING
On Launch
On Termination

34. Auto Scaling Lifecycle Hooks - Managing
aws autoscaling record-lifecycle-action-heartbeat
--lifecycle-action-token bcd2f1b8-9a78-44d3-8a7a-4dd07d7cf635
--lifecycle-hook-name my-launch-hook
--auto-scaling-group-name my-asg
aws autoscaling complete-lifecycle-action
--lifecycle-action-result CONTINUE
--lifecycle-action-token bcd2f1b8-9a78-44d3-8a7a-4dd07d7cf635
--lifecycle-hook-name my-launch-hook
--auto-scaling-group-name my-asg
Send Heartbeat to Extend
Complete Action Early

35. trainline

36. £2bn 28m 70%
visitors/moticket sales/yr mobile
trainline

37. Monolithic application
One release every 6 weeks
Legacy tech
Physical data centre
Centralised Ops
Environment Snowflakes

38. Vision
Go faster!
trainline

39. Organisation
Small Cross-
Functional
Teams
Functional
Teams
trainline

40. Architecture
Micro-servicesMonolith
trainline

41. Cloud Readiness
DNS
Tech Baseline
Singletons
Security
Licenses
Ops Ready
trainline

42. Top 10 AWS Services trainline
Dynamo DB LambdaAuto-ScalingEC2 S3
SQS SNS ECSCloudFormatio
n
Kinesis

43. Auto-Scaling Groups
Unit of Management
Scale-up Patching
Auto-Recovery
Lifecycle Hooks
trainline

44. Performance Tuning trainline
0.3s = £8m/year

45. Oracle Exadata
4TB, 600MB IO/sec
Latency critical
Target < 5% slower
16 core license
trainline

46. Oracle Exadata
AWS
Baseline
-25%
(32 core)
-12% -8% -5%
App
Tuning
trainline
Infra
Tuning
Database
Tuning
Today: +10%

47. Migration Approach
Big Bang
Database
Common
Infra Services Services
trainline

48. Continuous Delivery
Everything now Continuous Delivery
Downtime reduced 30%
trainline

49. Pets Cattle Baked Containers Lambda
Deployment Continuum trainline

50. Environment Manager
Self-service portal & API for common infrastructure
tasks
trainline
Environment Abstraction over AWS
Env & Infra Config Management
Deployment & Toggling
Scaling and Patching
Compare and Synchronise

51. Environment Manager trainline
1,300 servers actively managed
20,000 deployments since Jan
40% fewer Jira tickets
Improved visibility& productivity

59. Open Source!
(very soon)
trainline

60. 100% AWS
250 Micro-services
70+ Releases a Week
30% Less Downtime
2hrs for new Environment
£1.2m Annual Saving
trainline

61. trainline

62. Remember to complete
your evaluations!

63. Thank You