25 Best Practice Tips for architecting your Amazon VPC

According to me Amazon VPC is one of the most important feature introduced by AWS. We have been using AWS from 2008 and Amazon VPC from the day it was introduced and i strongly feel the customer adoption towards AWS cloud gained real momentum only after the introduction of VPC into the market.
Amazon VPC comes with lots of advantages over the limitations faced in Amazon Classic cloud like: Static private IP address , Elastic Network Interfaces :  possible to bind multiple Elastic Network Interfaces to a single instance, Internal Elastic Load Balancers, Advanced Network Access Control ,Setup a secure bastion host , DHCP options , Predictable internal IP ranges , Moving NICs and internal IPs between instances, VPN connectivity, Heightened security etc. Each and everything is a interesting topic on its own and i will be discussing them in detail in future.
Today i am sharing some of our implementation experience on working with hundreds of Amazon VPC deployments as best practice tips for the AWS user community. You can apply some of the relevant ones in your existing VPC or use these points as part of your migration approach to Amazon VPC.

Practice 1) Get your Amazon VPC combination right: Select the right Amazon VPC architecture first.  You need to decide the right Amazon VPC & VPN setup combination based on your current and future requirements. It is tough to modify/re-design the Amazon VPC at later stage, so it is better to design it taking into consideration your NW and expansion needs for next ~2 years. Currently different types of Amazon VPC setups are available; Like Public facing VPC, Public and Private setup VPC, Amazon VPC with Public and Private Subnets and Hardware VPN Access, Amazon VPC with Private Subnets and Hardware VPN Access, Software based VPN access etc. Choose the one which you feel you will be in next 1-2 years.

Practice 2) Choose your CIDR Blocks: While designing your Amazon VPC, the CIDR block should be chosen in consideration with the number of IP addresses needed and whether we are going to establish connectivity with our data center. The allowed block size is between a /28 netmask and /16 netmask. Amazon VPC can have contain from 16 to 65536 IP addresses. Currently Amazon VPC once created can’t be modified, so it is best to choose the CIDR block which has more IP addresses usually. Also when you design the Amazon VPC architecture to communicate with the on premise/data center ensure your CIDR range used in Amazon VPC does not overlaps or conflicts with the CIDR blocks in your On premise/Data center. Note: If you are using same CIDR blocks while configuring the customer gateway it may conflict.
E.g., Your VPC CIDR block is 10.0.0.0/16 and if you have 10.0.25.0/24 subnet in a data center the communication from instances in VPC to data center will not happen since the subnet is the part of the VPC CIDR. In order to avoid these consequences it is good to have the IP ranges in different class. Example., Amazon VPC is in 10.0.0.0/16 and data center is in 172.16.0.0/24 series.

Practice 3) Isolate according to your Use case: Create separate Amazon VPC for Development , Staging and Production environment (or) Create one Amazon VPC with Separate Subnets/Security/isolated NW groups for Production , Staging and development. We have observed 60% of the customer preferring the second choice. You chose the right one according to your use case.

Practice 4) Securing Amazon VPC : If you are running a machine critical workload demanding complex security needs you can secure the Amazon VPC like your on-premise data center or more sometimes. Some of the tips to secure your VPC are:

  • Secure your Amazon VPC using Firewall virtual appliance, Web application firewall available from Amazon Web Services Marketplace. You can use check point, Sophos etc for this
  • You can configure Intrusion Prevention or Intrusion Detection virtual appliances and secure the protocols and take preventive/corrective actions in your VPC
  • Configure VM encryption tools which encrypts your root and additional EBS volumes. The Key can be stored inside AWS (or) in your Data center outside Amazon Web Services depending on your compliance needs. http://harish11g.blogspot.in/2013/04/understanding-Amazon-Elastic-Block-Store-Securing-EBS-TrendMicro-SecureCloud.html
  • Configure Privileged Identity access management solutions on your Amazon VPC to monitor and audit the access of Administrators of your VPC.
  • Enable the cloud trail to audit in the VPC environments  ACL policy’s. Enable cloud trail :http://harish11g.blogspot.in/2014/01/Integrating-AWS-CloudTrail-with-Splunk-for-managed-services-monitoring-audit-compliance.html
  • Apply anti virus for cleansing specific EC2 instances inside VPC. Trend micro has very good product for this.
  • Configure Site to Site VPN for securely transferring information between Amazon VPC in different regions or between Amazon VPC to your On premise Data center
  • Follow the Security Groups and NW ACL’s best practices listed below

Practice 5) Understand Amazon VPC Limits: Always design the VPC subnets in consideration with the expansion in the future. Also understand the Amazon VPC’s limits before using the same. AWS has various limitations on the VPC components like Rules per security group, No of route tables and Subnets etc. Some of them may be increased after providing the request to the Amazon support team while few components cannot be increased. Ensure the limitations are not affecting your overall design. Refer URL:
http://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/VPC_Appendix_Limits.html

Practice 6) IAM your Amazon VPC: When you are going to assign people to maintain your Amazon VPC you can create Amazon IAM account with the fine grained permissions (or) use Sophisticated Privileged identity Management solutions available on AWS marketplace to IAM your VPC.

Practice 7) Disaster Recovery or Geo Distributed Amazon VPC Setup : When you are designing a Disaster Recovery Setup plan using VPC or expanding to another Amazon VPC region you can follow these simple rules. Create your Production site VPC CIDR : 10.0.0.0/16 and your DR region VPC CIDR:  172.16.0.0/16. Make sure they do not conflict with on premises subnet CIDR block in event both needs to be integrated to on premise DC as well. After CIDR blocks creation , setup a VPC tunnel between regions and to your on premise DC. This will help to replicate your data using private IP’s.

Practice 8) Use security groups and Network ACLs wisely:  It is advisable to use security groups over Network ACLs inside Amazon VPC wherever applicable for better control. Security groups are applicable on EC2 instance level while network ACL is applicable on Subnet level.  Security groups are used for White list mostly. To blacklist IPs, one can use Network ACLs.

Practice 9) Tier your Security Groups : Create different security groups for different tiers of your infrastructure architecture inside your VPC. If you have Web, App, DB tiers create different security group for each of them. Creating tier wise security groups will increase the infrastructure security inside Amazon VPC.  EC2 instances in each tier can talk only on application specified ports and not at all ports. If you create Amazon VPC security groups for each and every tier/service separately it will be easier to open a port to a particular service. Don’t use same security group for multiple tiers of instances, this is a bad practice.
Example: Open ports for security group instead of IP ranges : For example : People have tendency to open for port 8080 to 10.10.0.0/24 (web layer) range. Instead of that, open port 8080 to web-security-group. This will make sure only web security group instances will be able to contact on port 8080. If someone launches NAT instance with NAT-Security-Group in 10.10.0.0/24, he won’t be able to contact on port 8080 as it allows access from only web security group.
Practice 10 ) Standardize your Security Group Naming conventions : Following a security group naming conventions inside Amazon VPC will improve operations/management for large scale deployments inside VPC. It also avoids manual errors, leaks and saves cost and time overall.
For example: Simple ones like Prod_DMZ_Web_SG or Dev_MGMT_Utility_SG (or) complex coded ones for large scale deployments like
USVA5LXWEBP001- US East Virginia AZ 5 Linux Web Server Production 001
This helps in better management of security groups.
Practice 11) ELB on Amazon VPC:  When using Amazon ELB for Web Applications, put all other EC2 instances( Tiers like App,cache,DB,BG etc)  in private subnets as much possible. Unless there is a specific requirement where instances need outside world access and EIP attached, put all instances in private subnet only. Only ELBs should be provisioned in Public Subnet as secure practice in Amazon VPC environment.
Practice 12) Control your outgoing traffic in Amazon VPC: If you are looking for better security, for the traffic going to internet gateway use Software’s like Squid or Sophos to restrict the ports,URL,Domains etc so that all traffic go through the proxy tier controlled and it also gets logged. Using these proxy/security systems we can also restrict the unwanted ports, by doing so,  if there is any security compromise to the application running inside Amazon VPC they can be detected by auditing the restricted connections captured from the logs. This helps in corrective security measure.
Practice 13) Plan your NAT Instance Type: Whenever your Application EC2 instances residing inside private subnet of Amazon VPC are making Web Service/HTTP/S3/SQS calls they go through NAT instance. If you have designed Auto scaling for your application tier and there are chances ten’s of app EC2 instances are going to make lots of web calls concurrently, NAT instance will become a performance bottleneck at this juncture. Size your NAT instance capacity depending upon application needs for avoiding performance bottlenecks. Using the NAT instances provides us with advantages of saving cost of Elastic IP and provides extra security by not exposing the instances to outside world for accessing the internet.
Practice 14) Spread your NAT instance with Multiple Subnets: What if you have hundreds of EC2 instances inside your Amazon VPC and they are making lots of heavy web service/HTTP calls concurrently. A single NAT instance with even largest EC2 size cannot handle that bandwidth sometimes and may become performance bottleneck. In Such scenarios, span your EC2 across multiple subnets and create NAT’s for each subnet. This way you can spread your out going bandwidth and improve the performance in your VPC based deployments.
Practice 15) Use EIP when needed: At times you may need to keep a part of your application services to be kept in Public subnet for external communication. It is recommended practice to associate them with Amazon Elastic IP and white list these IP address in the target services used by them
Practice 16) NAT instance practices : If needed, enable Multi factor authentication on NAT instance. SSH and RDP ports are open only on sources and destination IP’s, not global network (0.0.0.0/0). SSH / RDP ports are opened only on static exit IP’s not dynamic exit IP’s.
Practice 17) Plan your Tunnel between On-Premise DC to Amazon VPC: 
Select the right mechanism to connect your on premises DC to Amazon VPC. This will help you to connect the EC2 instance via private IP’s in a secure manner.
  • Option 1: Secure IPSec tunnel to connect a corporate network with Amazon VPC (http://aws.amazon.com/articles/8800869755706543)
  • Option 2 : Secure communication between sites using the AWS VPN CloudHub (http://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/VPN_CloudHub.html)
  • Option 3: Use Direct connect between Amazon VPC and on premise when you have lots of data to be transferred with reduced latency (or) you have spread your mission critical workloads across cloud and on premise. Example: Oracle RAC in your DC and Web/App tier in your Amazon VPC. Contact us if you need help on setting up direct connect between Amazon VPC and DC.
Practice 18) Always span your Amazon VPC across multiple subnets in Multiple Availability zones inside a Region. This helps is architecting high availability inside your Amazon VPC properly. Example: Classification of the VPC subnet : WEB Tier Subnet : 10.0.10.0/24 in Az1 and 10.0.11.0/24 in Az2, Application Tier Subnet :  10.0.12.0/24 and 10.0.13.0/24, DB Tier Subnet :  10.0.14.0/24 and 10.0.15.0/24, Cache Tier Subnet : 10.0.16.0/24 and 10.0.17.0/24 etc
Practice 19) Good security practice is that to have only public subnet with route table which carries route to internet gateway. Apply this wherever applicable.
Practice 20) Keep your Data closer : For small scale deployments in AWS where cost is critical than high availability, It is better to keep the Web/App in same availability zone as of ElastiCache , RDS etc inside your Amazon VPC. Design your subnets accordingly to suit this. This is not a recommended architecture for applications demanding High Availability.
Practice 21) Allow and Deny Network ACL : Create Internet outbound allow and deny network ACL in your VPC.
First network ACL: Allow all the HTTP and HTTPS outbound traffic on public internet facing subnet.
Second network ACL: Deny all the HTTP/HTTPS traffic. Allow all the traffic to Squid proxy server or any virtual appliance.
Practice 22 ) Restricting Network ACL : Block all the inbound and outbound ports. Only allow application request ports. These are stateless traffic filters that apply to all traffic inbound or outbound from a Subnet within VPC. AWS recommended Outbound rules : http://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/VPC_Appendix_NACLs.html
Practice 23) Create route tables only when needed and use the Associations option to map subnets to the route table in your Amazon VPC
Practice 24) Use Amazon VPC Peering (new) : Amazon Web Services has introduced VPC peering feature which is quite useful one. AWS VPC peering connection is a networking connection between two Amazon VPCs that enables you to route traffic between them using private IP addresses. Currently it can be in same AWS region, Instances in either VPC can communicate with each other as if they are within the same network. Since AWS uses the existing infrastructure of a VPC to create a VPC peering connection; it is neither a gateway nor a VPN connection, and does not rely on a separate piece of physical hardware (which essentially means there is no single point of failure for communication or a bandwidth bottleneck).

We have seen it is useful in following scenarios :
  1. Large Enterprises usually run Multiple Amazon VPC in single region and some of their applications are so interconnected that they may need to access them privately + securely inside AWS. Example Active Directory, Exchange, Common business services will be usually interconnected.
  2. Large Enterprise have different AWS accounts for different business units/teams/departments , at times systems deployed by some business units in different AWS accounts need to be shared or need to consume a shared resource privately. Example: CRM , HRMS ,File Sharing etc can be internal and shared. In such scenarios VPC peering comes very useful.
  3. Customer can peer their VPC with their core suppliers to have tighter integrated access of their systems.
  4. Companies offering Infra/Application Managed Services on AWS can now safely peer into customer Amazon VPC and provide monitoring and management of AWS resources.

Practice 25) Use Amazon VPC: It is highly recommended that migrate all your new workloads inside Amazon VPC rather than Amazon Classic Cloud. I also strongly recommend to migrate your existing workloads from Amazon Classic cloud to Amazon VPC in phases or one shot which ever is feasible. In addition to the benefits of the VPC that is detailed in the start of the article, AWS has started introducing lots of features which are compatible only inside VPC and in the AWS marketplace as well there are lots of products which are compatible only with Amazon VPC.  So make sure you leverage this strength of VPC. If you require any help for this migration please contact me.

readers feel free to suggest more.. I will link relevant ones in this article

Architecting Highly Available ElastiCache Redis replication cluster in AWS VPC

In this post lets explore how to architect and create a Highly Available + Scalable Redis Cache Cluster for your web application in AWS VPC. Following is the architecture in which the ElastiCache Redis Cluster is assembled:

  • Redis Cache Cluster inside Amazon VPC for better control and security
  • Master Redis Node 1 will be created in AZ-1 of US-West
  • Redis Read Replica Node 2 will be created in AZ-2 of US-West
  • Redis Read Replica Node 3 will be created in AZ-3 of US-West

You can position all the 3 Redis Nodes in different Availability zones for Achieving High Availability (or) you can position Master + RR 1 in AZ1 and RR 2 in AZ2. This reduces the Inter – AZ latency and might give better performance for heavily used clusters.
Step 1: Creating Cache Subnet groups:
To create Cache Subnet group  navigate to the dashboard of ElastiCache, select Cache Subnet groups and then click “Create Cache Subnet group”. Add the Subnet Id and the Availability Zone you need to use for the ElastiCache cluster.

 

We have created Amazon VPC spreading across 3 availability zones. In this post we are going to place the Redis Master and 2 Redis Replica Slaves in these 3 availability zones. Since Redis will be most of the times accessed by your application tier it is better if you place them on Private Subnet of your VPC.
Step 2: Creating Redis Cache Cluster: 
To create Cache Cluster navigate to the  dashboard of ElastiCache, select Launch Cache Cluster and provide the necessary details. We are launching it inside Amazon VPC, so we have to select the Cache Subnet group .
Note: It is mandatory to create Cache Subnet group before Launch if you need ElastiCache Redis cluster in Amazon VPC.

 

For test purposes i have used m1.small EC2 instance for the Redis. Since this is a fresh Redis installation, i have not mentioned S3 bucket from where the persistent Redis Snapshot will be used as input. On successful creation of the Cache Cluster you can see the details in the dashboard.
Step 3: Replication Group Creation:
To create Replication group select the option of Replication Groups from dashboard and then select the “Create Replication Group”

Select the master Redis node “redisinsidevpc” created previously as the primary cluster id of the Cache cluster.  Give the Replication group id and description as illustrated below.

Note: Replication Group should be created only after the Primary Cache Cluster node is UP and running, else you will get the error as shown below.

On the successful creation of the Replication group you can see the following details. You can observe from below screenshot that there is only one primary node in US-WEST-2A and zero Redis Read Replica’s are attached to it.

Step 4: Adding Read Replica Nodes:
When you select the Replication group, you can see the option to add Redis Read Replica. We are adding 2 Redis Read Replica named Redis-RR1 (in US-West-2B) and Redis-RR2 (in US-WEST-2C). Both the Read replica’s are pointed to the master node “redisinsidevpc”. Currently we can add up to 5 Read replica Nodes for a Redis Master Node. This is more than enough to handle Thousands of messages per second. If you combine it with Redis Pipeline handling 100K messages per second from a node is like cake walk.
Adding Read Replica 1 in Us-West -2B

Adding Read Replica 2 in US-West-2c

On successful creation you can see the following details of Replication group in the dashboard. Now you can see there are 3 Redis nodes listed with Number of read Replica’s as 2. Placing the Read Replica’s and master node in multiple AZ will increase the high availability and protects you from node and AZ level failure. On our sample tests inter AZ Replication deployments had <2 second replication lag for massive writes on master and <1 second replication lag between master slave inside same AZ deployments. We pumped @100K messages per second for few minutes on m1.large Redis instance cluster.
In event, if you need additional read scalability i recommend to use more read Replica slaves added to the master.
In your application tier you need to use the primary Endpoint “redis-replication.qcdze2.0001.usw2.cache.amazon.aws.com:6379” shown below to connect to Redis.

If you need to delete/reboot/Modify you can make it through the options available here.

Step 5: Promoting the Read replica:

You can also promote any node as the Primary cluster using the Promote/Demote option. There will be only one Primary Node.
Note: This step is not part of the cluster creation process.

This promotion has to be carried out with caution and proper understanding for maintaining data consistency.

Post was co authored with Senthil 8KMiles

Billion Messages – Art of Architecting scalable ElastiCache Redis tier`

Whenever we are designing a highly scalable architectures on AWS running thousands of application servers and supporting millions of requests, usage of NoSQL solutions have become inevitable part. One such solution we often been using for years on AWS is Redis . We love Redis. 
AWS introduced ElastiCache Redis on 2013 and we started using the same since it eased the management and operational efforts.  In this article i am going to share my experience on designing large scale Redis tiers supporting billions of messages per day on AWS, step by step guide on how to deploy the same, what are the Implications you face at scale ? Best Practices to be adopted while designing sharded+replicated Redis Tiers etc.

Since we need to support billions of message requests per day and it was growing:

  • the ElastiCache Redis tier was designed with Partitions( shards) to scale out as the customer grows
  • the ElastiCache Redis tier was designed with Replica Slaves for HA and read scaling as the read volumes grow

When your application is growing at Rapid pace and lots of data are created every day, you cannot keep increasing (scaling up) the size of the ElastiCache Node. At one point you will hit the maximum memory capacity of your EC2 instance and you will be forced to partition.  Partitioning is the process of splitting your Key Value data into multiple ElastiCache Redis instances, so that every instance will only contain a subset of your Key Value pair. It allows for much larger ElastiCache Redis data stores, using the sum of the memory of many ElastiCache Redis Nodes. It also allows to scale the computational power to multiple cores and multiple EC2, and the network bandwidth to multiple EC2 network adapters. There are two widely used partition/shard implementation techniques that are available for ElastiCache Redis Tier :
Technique 1) Client side partitioning means that the Redis clients directly select the right ElastiCache Redis node where to write or read a given key. Many Redis clients implement client side partitioning, chose the right one wisely.
Technique 2) Proxy assisted partitioning means that your clients send requests to a proxy that is able to speak the Redis protocol, which in turn sends requests directly to the right ElastiCache Redis instance. The proxy will make sure to forward our request to the right Redis instance accordingly to the configured partitioning schema. Currently the most widely used Proxy assisted partitioning tool is Twemproxy , written by Manju Raj of twitter. Git hub link https://github.com/twitter/twemproxy . Twemproxy is a proxy developed at Twitter for the Memcached ASCII and the Redis protocol. Twemproxy supports automatic partitioning among multiple Redis instances and  currently it is the suggested way to handle partitioning with Redis.

In this article we are going to explore in detail about Proxy assisted partitioning technique for highly scalable and available Redis tier.

Welcome to Twemproxy

Twemproxy( nutcracker) is a fast single-threaded proxy supporting the Memcached ASCII protocol and more recently the Redis protocol.

Installing Twemproxy:

Download the Twemproxy package.
wget http://twemproxy.googlecode.com/files/nutcracker-0.3.0.tar.gz
tar -xf nutcracker-0.3.0.tar.gz
cd nutcracker-0.3.0
./configure
make
make install

Configuration:

Twemproxy (Nutcracker) can be configured through a YAML file specified by the -c or –conf-file command-line argument on process start. The configuration file is used to specify the server pools and the servers within each pool that nutcracker manages. The configuration files parses and understands the following keys:

• listen: The listening address and port (name:port or ip:port) for this server pool.
• hash: The name of the hash function.
• hash_tag: A two character string that specifies the part of the key used for hashing. Eg “{}” or “$$”. Hash tag enable mapping different keys to the same server as long as the part of the key within the tag is the same.
• distribution: The key distribution mode.
• timeout: The timeout value in msec that we wait for to establish a connection to the server or receive a response from a server. By default, we wait indefinitely.
• backlog: The TCP backlog argument. Defaults to 512.
• preconnect: A boolean value that controls if nutcracker should preconnect to all the servers in this pool on process start. Defaults to false.
• redis: A boolean value that controls if a server pool speaks redis or memcached protocol. Defaults to false.
• server_connections: The maximum number of connections that can be opened to each server. By default, we open at most 1 server connection.
• auto_eject_hosts: A boolean value that controls if server should be ejected temporarily when it fails consecutively server_failure_limit times. See liveness recommendations for information. Defaults to false.
• server_retry_timeout: The timeout value in msec to wait for before retrying on a temporarily ejected server, when auto_eject_host is set to true. Defaults to 30000 msec.
• server_failure_limit: The number of consecutive failures on a server that would lead to it being temporarily ejected when auto_eject_host is set to true. Defaults to 2.
• servers: A list of server address, port and weight (name:port:weight or ip:port:weight) for this server pool.

For More details Refer: https://github.com/twitter/twemproxy

Running and Accessing Twemproxy 

To start the proxy just use the command “nutcracker” with the configuration file path specified or in its default path(conf/nutcracker.yml) .
Based on the configuration the twemproxy will be running and listening. Configure your application to point to the port and address instead of the Redis cluster.

Twemproxy Deployment models:

We usually deploy Twemproxy one one of the following models in AWS :

Model 1: Twemproxy as a separate Proxy Tier: In this model Twemproxies are deployed in separate EC2 instances, The application tier is configured to point to Twemproxies . The Twemproxy tier in turn maintains the mappings to the ElastiCache redis nodes. It is better to use instances with very good IO bandwidth for twemproxy tier in AWS. In case you feel the instance CPU is underutilized, you can launch multiple Twemproxy instances inside the same single EC2 instance as well.

Though the above model looks clean and efficient there are optimizations that can be applied to this architecture :
What happens when the twemproxy01 fails, how will the Application server instances know about it ?
Why should i pay additional for twemproxy EC2 instances, Can it be minimized ?

Model 2 : Twemproxy bundled with application tier EC2’s: 

In this model twemproxies are bundled in the same box of the application server EC2 itself. Since two twemproxies are not aware of each others existence, it is easy to architect this model even in App->Auto Scaling mode. Every application server talks to the local twemproxy deployed in the same box this saves cost and avoids managing additional tier complexity as well.

Reference ElastiCache Redis + Twemproxy  deployment:

(This is a Reference deployment, the same can be scaled out to hundreds depending upon the need. It is a Redis Partitioned + replicated setup )
1. Two ElastiCache Redis nodes in AWS (twem01 and twem02)
2. Replication group for each ElastiCache redis nodes (twem01-rg and twem02-rg with one Read Replica each)
3. Two twemproxy servers running in separate EC2. (twemproxy01 and twemproxy02)
Once the above setup is done please note down the endpoints. We will be using the Replication group endpoint as the ElastiCache Redis endpoint for the twemproxy.

ElastiCache Redis Endpoints:

twem01-twem01.qcdze2.0001.usw2.cache.amazonaws.com:6379
twem02-twem02.qcdze2.0001.usw2.cache.amazonaws.com:6379
ElastiCache Redis Replication endpoints:

twem01-rg.qcdze2.ng.0001.usw2.cache.amazonaws.com:6379
twem02-rg.qcdze2.ng.0001.usw2.cache.amazonaws.com:6379

To test the Twemproxy we pumped following keys:
Pump KV data through the Twemproxy01 (1-2000 keys)
Pump KV data through the Twemproxy02(2001-4000 keys).

Configuration:
beta:
listen: 127.0.0.1:22122
hash: fnv1a_64
hash_tag: “{}”
distribution: ketama   #Consistent Hashing
auto_eject_hosts: false
timeout: 5000
redis: true
servers:
– twem01-rg.qcdze2.ng.0001.usw2.cache.amazonaws.com:6379:1 server1
– twem02-rg.qcdze2.ng.0001.usw2.cache.amazonaws.com:6379:1 server2

Test 1: Testing Key accessibility . Testing “GET” operation across both the Twemproxy Instances for few sample keys. 

Fetch 4 Keys spread across 4000 KV data from Twemproxy01  EC2 instance:
[root@twemproxy01 redish]# src/redis-cli -h 127.0.0.1 -p 22122
redis 127.0.0.1:22122> get 1000
“1000-data”
redis 127.0.0.1:22122> get 2000
“2000-data”
redis 127.0.0.1:22122> get 3000
“3000-data”
redis 127.0.0.1:22122> get 4000
“4000-data”
Fetch 4 Keys spread across 4000 KV data from Twemproxy02  EC2 instance:
[root@twemproxy02 redish]# src/redis-cli -h 127.0.0.1 -p 22122
redis 127.0.0.1:22122> get 1000
“1000-data”
redis 127.0.0.1:22122> get 2000
“2000-data”
redis 127.0.0.1:22122> get 3000
“3000-data”
redis 127.0.0.1:22122> get 4000
“4000-data”

From the above test it is evident that all 4000 KV data inserted using both Twemproxies are accessible from both Twemproxies( testing the sample) even though they are not aware among themselves. This is because of the same hashing and Key mapping translation done at Twemproxy level.

Test 2: Testing the ElastiCache Redis Availability and Fail over mechanism:

We are going to promote the twem01-rg replication group read replica to be the Primary Redis Node. After promotion we are going to test:

 

  1. Whether the Twemproxy is able to recognize the newly promoted master
  2. Whether the sample KV data is safely replicated and still accessible , to ensure failover is successful.

To promote ElastiCache Redis slave just click the promote Action and confirm or automate using API. During the promotion of Read Replica to master we observed that the transition happens very quickly and there is no timeout but the response time for the query is about 4-5 secs for about 3-4 minutes during the switch over. In the Twemproxy configuration we can set the timeout configuration, this value needs to be set accordingly so that during switch over there will be no connection refused. For the sample test we have set it as 5000

Repeat Test 1:

[root@twemproxy01 redish]# src/redis-cli -h 127.0.0.1 -p 22122
redis 127.0.0.1:22122> get 1000
“1000-data”
redis 127.0.0.1:22122> get 2000
“2000-data”
redis 127.0.0.1:22122> get 3000
“3000-data”
redis 127.0.0.1:22122> get 4000
“4000-data”
Fetch 4 Keys spread across 4000 KV data from Twemproxy02  EC2 instance:
[root@twemproxy02 redish]# src/redis-cli -h 127.0.0.1 -p 22122
redis 127.0.0.1:22122> get 1000
“1000-data”
redis 127.0.0.1:22122> get 2000
“2000-data”
redis 127.0.0.1:22122> get 3000
“3000-data”
redis 127.0.0.1:22122> get 4000
“4000-data”

From the above test it is evident that all 4000 KV data are replicated properly between master and slaves nodes and the transition between slave to master happened successfully with all the data.
Reporting

Nutcracker exposes stats at the granularity of server pool and servers per pool through the stats monitoring port. The stats are essentially JSON formatted key-value pairs, with the keys corresponding to counter names. By default stats are exposed on port 22222 and aggregated every 30 seconds.

Some best practices while designing highly scalable+available ElastiCache Redis Tier :

Practice 1 : Reduce the Number of Connections and pipeline messages:

Whenever the application instance gets a request to get/put value to the ElastiCache redis node, the client makes a connection to the Redis Tier. Imagine it is a heavy traffic site, then thousands of requests hitting translates to thousands of connections from the application instance to Redis Tier. Now when you add Auto- scaling to your application tier and you have few hundred servers scaled out , then imagine the connection complexity and overhead this architecture brings to the ElastiCache Redis Tier.

Best practice is minimize the number of connections made from your application instance to your ElastiCache redis node. Use Twemproxy in bundled mode with Application EC2 instance, this keeps the process in close proximity and reduces the connection overhead.  Secondly, Twemproxy internally uses minimal connections to ElastiCache Redis Instance by proxying multiple client connections onto one or few server connections.
Redis also supports pipelines, where multiple requests can be pipelined and sent on a single connection. In a simple test using large Application & ElastiCache node we were able to process 125K message/sec in pipeline mode, now imagine what you could achieve on bigger instance types on AWS. The connection minimization architectural setup of twemproxy makes it ideal for pipelining requests and responses and hence saving on the round trip time.  For example, if twemproxy is proxying three client connections onto a single server and we get requests – ‘get key\r\n’, ‘set key 0 0 3\r\nval\r\n’ and ‘delete key\r\n’ on these three connections respectively, twemproxy would try to batch these requests and send them as a single message onto the server connection.

Note : It is important to note that “read my last write” constraint doesn’t necessarily hold true when twemproxy is configured withserver_connections: > 1. Let us consider a scenario where twemproxy is configured with server_connections: 2. If a client makes pipelined requests with the first request in pipeline being set foo 0 0 3\r\nbar\r\n (write) and the second request being get foo\r\n (read), the expectation is that the read of key foo would return the value bar. However, with configuration of two server connections it is possible that write and read request are sent on different server connections which would mean that their completion could race with one another. In summary, if the client expects “read my last write” constraint, you either configure twemproxy to use server_connections:1 or use clients that only make synchronous requests to twemproxy.

Practice 2:  Configure Auto Ejection and Hashing combination properly

Design for failure is the mantra of cloud architecture. Failures are commons when things are distributed on scale. Though partitioning when using ElastiCache Redis as a data store or cache is conceptually the same on broad lines, there is a huge difference operationally on large scale systems. When you are using ElastiCache Redis as a data store you need to be sure that a given key always maps to the same instance, Whereas if you are using  ElastiCache Redis as cache if a given node is not available, then you can always start afresh using a different node in the hash ring with consistent hashing implementations.
To be resilient against failures, it is recommended that you configure Auto eject hosts false when you treat redis as a Data Store and true in when you treat redis as a cache.
resilient_pool:
auto_eject_hosts: true
server_retry_timeout: 30000
server_failure_limit: 3
Enabling auto_eject_hosts: This property ensures that a dead ElastiCache redis Node can be ejected out of the hash ring after server_failure_limit: consecutive failures have been encountered on that node. A non-zero server_retry_timeout: ensures that we don’t incorrectly mark a node as dead forever especially when the failures were really transient. The combination of server_retry_timeout: and server_failure_limit: controls the tradeoff between resiliency to permanent and transient failures.
Note that an ejected node will not be included in the hash ring for any requests until the retry timeout passes. This will lead to data partitioning as keys originally on the ejected node will now be written to another node still in the pool. If ElastiCache Redis is used as a cache (in memory) then in event of a Redis Node going down, the cache data will be lost. This cache miss can cascade performance problems to other tiers and altogether bring down your system on the cloud. To minimize KV cache miss,  you can design your hash ring with Ketama hashing on the Redis Proxy. This will minimize the Cache miss in event of cache node failure, also it decreases the overall re-balancing needed in your Redis tier.  In addition to helping hand on availability problems, Redis Proxy+Ketama can also help your Redis farm to Scale out and Scale down easily with minimal cache miss. To know more about Ketama on ElastiCache refer http://harish11g.blogspot.com/2013/01/amazon-elasticache-memcached-internals_8.html  .
The below diagram illustrates a ElastiCache Redis Cache Farm with Consistent Hash Ring.
In short to minimize the cache miss when using auto eject with true it is recommended to use “Ketama Hashing ( Consistent Hashing Algorithm)” on your Twemproxy configuration. 
ElastiCache Redis as a Data Store:

What if the data stored in your Cache is important and needs to persisted across node failures and launch ? What if the date stored in your Cache cannot be lost and it needs to be replicated and promoted during failures?
Welcome to ElastiCache Redis as Data store. ElastiCache Redis offers features to persist the in memory cache data to disk and also replicate it to a slave for high availability. If ElastiCache Redis is used as a store (persistent), you need to keep the map between keys and nodes fixed, and a fixed number of nodes. Since the data stored is important when you treat ElastiCache Redis as a data store, in event one Redis node goes down, you should have immediate standby up and running in minutes.  You can architect ElastiCache Redis master with one or more replication Slave launched on different AZ from Master for High Availability in AWS. In event master node failure or master AZ failure, the slave Redis node can be promoted in minutes to act as master. This whole High availability design keeps the number of nodes on the hash ring stable and simple, Otherwise, you will end up building a system to re balance the keys (which is not easy) between nodes whenever there is a addition or removal of nodes during outages. In addition to above the ElastiCache Redis supports Partial Resynchronization with Slaves – If the connection between a master node and a slave node is momentarily broken, the master now accumulates data that is destined for the slave in a backlog buffer. If the connection is restored before the buffer becomes full, a quick partial resync will be done instead of a potentially longer full resync. This really saves network bottleneck during momentary failures.
In large scale systems you will often find some partitions are heavily used than others , in event the usage is read heavy in nature you can add upto 5 Read replicas for the ElastiCache Redis Master partition. Since these replicas are used only for read they do not affect the Hash ring structure. But Twemproxy lacks the support for read scaling with Redis Replica’s. So in event when you face this problem, you will have to Scale up the capacity(instance/node type) of the Master and Slave of that partition alone.

If you are using ElastiCache redis as a Data store in the TwemProxy it is recommended to keep “auto_eject_hosts” property false so that in event of redis node failure it is not ejected from the hash ring. The hash ring can be built with both ketama or modula hash algorithms , since in event of Primary node failure, the Slave is going to be promoted and ring structure is going to be always maintained. But if you feel there is immense possibility for the number of primary node partitions to grow, or major failures to occu, it is better to choose ketama hash ring itself from beginning. The below diagram illustrates the architecture.

Practice 3: Configure the Buffer properly:

All memory for incoming requests and outgoing responses is allocated in mbuf in Twemproxy. Mbuf enables zero copy for requests and responses flowing through the proxy. By default an mbuf is 16K bytes in size and this value can be tuned between 512 and 16M bytes using -m or –mbuf-size=N argument. Every connection has at least one mbuf allocated to it. This means that the number of concurrent connections twemproxy can support is dependent on the mbuf size. A small mbuf allows us to handle more connections, while a large mbuf allows us to read and write more data to and from kernel socket buffers. Large Scale web/mobile applications involving millions of hits might have small size request/response and lots of concurrent connections to handle in their backend. So at such scenarios, when Twemproxy is meant to handle a large number of concurrent client connections, you should set chunk size to a small value like 512 bytes to 1K bytes using the -m or –mbuf-size=N argument.

Practice 4: Configure proper Timeouts
It is always a good idea to configure Twemproxy timeout: for every server pool, rather than purely relying on client-side timeouts. Eg:

resilient_pool_with_timeout:
auto_eject_hosts: true
server_retry_timeout: 30000
server_failure_limit: 3
timeout: 400
Relying only on client-side timeouts has the adverse effect of the original request having timed out on the client to proxy connection, but still pending and outstanding on the proxy to server connection. This further gets exacerbated when client retries the original request.

Benefits of using Twemproxy for Redis Scaling

  • Avoids re inventing the wheel. Thanks Manju Raj (twitter).
  • reduce the number of connections to your cache server by acting as a proxy
  • shard data automatically between multiple cache servers
  • support consistent hashing with different strategies and hashing functions
  • be configured to disable nodes on failure
  • run in multiple instances, allowing client to connect to the first available proxy server
  • Pipelining and batching of requests and hence saving of round-trips

Disadvantages of Partitioning Model:

Point 1) Operations involving multiple keys are usually not supported. For instance you can’t perform the intersection between two sets if they are stored in keys that are mapped to different Redis instances (actually there are ways to do this, but not directly).Redis transactions involving multiple keys can not be used.
Point 2) The partitioning granularity is the key, so it is not possible to shard a dataset with a single huge key like a very big sorted set. Ideally in such cases you should Scale UP the particular Redis Master-Slave to larger EC2 instance or pro grammatically stitch up the sorted set.
Point 3)When partitioning is used, data handling is more complex, for instance you have to handle multiple RDB / AOF files, and to make a backup of your data you need to aggregate the persistence files/snapshots from multiple EC2 Redis slaves.
Point 4) Architecting a partitioned + replicated ElastiCache Redis tier not complex. What is more complex is ? supporting transparent rebalancing of data with the ability to add and remove nodes at runtime. Systems like client side partitioning and proxies don’t support this feature. However a technique called Presharding helps in this regard with limitations. Presharding technique ->Since Redis is lightweight, you can start with a lot of EC2 instances since the beginning itself. For example if you start with 32 or 64 EC2 instances (micro or small Cache Node instance type)  as your node capacity , it will provide enough room to keep scaling up the capacity when your data storage needs increase. It is not a highly recommended technique. But still can be used in production if your growth pattern is very predictable.

Future of highly scalable + available Redis tiers -> Redis Cluster

Redis Cluster is the preferred way to get automatic sharding and high availability. It is currently not production ready. Once Redis Cluster / Client  is available on Amazon ElastiCache, it will be the de facto standard for Redis partitioning. It uses a mix between query routing and client side partitioning.

References:
http://redis.io/documentation
https://github.com/twitter/twemproxy

This article was co-authored with Senthil