Intelligent People Uncommon Ideas Building a Scalable Architecture

Intelligent People. Uncommon Ideas. Building a Scalable Architecture for Web Apps Part I (Lessons Learned @ Directi) By Bhavin Turakhia CEO, Directi (http: //www. directi. com | http: //wiki. directi. com | http: //careers. directi. com) Licensed under Creative Commons Attribution Sharealike Noncommercial 1

Agenda • • • Why is Scalability important Introduction to the Variables and Factors Building our own Scalable Architecture (in incremental steps) § § § • • Vertical Scaling Vertical Partitioning Horizontal Scaling Horizontal Partitioning … etc Platform Selection Considerations Tips Creative Commons Sharealike Attributions Noncommercial 2

Why is Scalability Important in a Web 2. 0 world • • Viral marketing can result in instant successes RSS / Ajax / SOA § pull based / polling type § XML protocols - Meta-data > data § Number of Requests exponentially grows with user base • • Ro. R / Grails – Dynamic language landscape gaining popularity In the end you want to build a Web 2. 0 app that can serve millions of users with ZERO downtime Creative Commons Sharealike Attributions Noncommercial 3

The Variables • Scalability - Number of users / sessions / transactions / operations the entire system can perform • • • Performance – Optimal utilization of resources Responsiveness – Time taken per operation Availability - Probability of the application or a portion of the application being available at any given point in time • Downtime Impact - The impact of a downtime of a server/service/resource - number of users, type of impact etc • • Cost Maintenance Effort High: scalability, availability, performance & responsiveness Low: downtime impact, cost & maintenance effort Creative Commons Sharealike Attributions Noncommercial 4

The Factors • • • Platform selection Hardware Application Design Database/Datastore Structure and Architecture Deployment Architecture Storage Architecture Abuse prevention Monitoring mechanisms … and more Creative Commons Sharealike Attributions Noncommercial 5

Lets Start … • We will now build an example architecture for an example app using the following iterative incremental steps – § § § Inspect current Architecture Identify Scalability Bottlenecks Identify SPOFs and Availability Issues Identify Downtime Impact Risk Zones Apply one of • • Vertical Scaling Vertical Partitioning Horizontal Scaling Horizontal Partitioning § Repeat process Creative Commons Sharealike Attributions Noncommercial 6

Step 1 – Lets Start … Appserver & DBServer Creative Commons Sharealike Attributions Noncommercial 7

Step 2 – Vertical Scaling Appserver, DBServer CPU RAM Creative Commons Sharealike Attributions Noncommercial 8

Step 2 - Vertical Scaling • Appserver, DBServer CPU CPU RAM RAM • • Introduction § Increasing the hardware resources without changing the number of nodes § Referred to as “Scaling up” the Server Advantages § Simple to implement Disadvantages § Finite limit § Hardware does not scale linearly (diminishing returns for each incremental unit) § Requires downtime § Increases Downtime Impact § Incremental costs increase exponentially Creative Commons Sharealike Attributions Noncommercial 9

Step 3 – Vertical Partitioning (Services) • Introduction § Deploying each service on a separate node • Positives § Increases per application Availability § Task-based specialization, optimization and tuning possible § Reduces context switching § Simple to implement for out of band processes § No changes to App required § Flexibility increases App. Server DBServer • Negatives § Sub-optimal resource utilization § May not increase overall availability § Finite Scalability Creative Commons Sharealike Attributions Noncommercial 10

Understanding Vertical Partitioning • The term Vertical Partitioning denotes – § Increase in the number of nodes by distributing the tasks/functions § Each node (or cluster) performs separate Tasks § Each node (or cluster) is different from the other • Vertical Partitioning can be performed at various layers (App / Server / Data / Hardware etc) Creative Commons Sharealike Attributions Noncommercial 11

Step 4 – Horizontal Scaling (App Server) • Load Balancer App. Server Introduction § Increasing the number of nodes of the App Server through Load Balancing § Referred to as “Scaling out” the App Server DBServer Creative Commons Sharealike Attributions Noncommercial 12

Understanding Horizontal Scaling • The term Horizontal Scaling denotes – § § Increase in the number of nodes by replicating the nodes Each node performs the same Tasks Each node is identical Typically the collection of nodes maybe known as a cluster (though the term cluster is often misused) § Also referred to as “Scaling Out” • Horizontal Scaling can be performed for any particular type of node (App. Server / DBServer etc) Creative Commons Sharealike Attributions Noncommercial 13

Load Balancer – Hardware vs Software • • • Hardware Load balancers are faster Software Load balancers are more customizable With HTTP Servers load balancing is typically combined with http accelerators Creative Commons Sharealike Attributions Noncommercial 14

Load Balancer – Session Management • Sticky Sessions § Requests for a given user are sent to a fixed App Server § Observations • Asymmetrical load distribution Sticky Sessions User 1 (especially during downtimes) • Downtime Impact – Loss of session data User 2 Load Balancer App. Server Creative Commons Sharealike Attributions Noncommercial App. Server 15

Load Balancer – Session Management • Central Session Store § Introduces SPOF § An additional variable § Session reads and writes generate Disk + Network I/O § Also known as a Shared Session Store Cluster Central Session Storage Load Balancer App. Server Session Store Creative Commons Sharealike Attributions Noncommercial 16

Load Balancer – Session Management • Clustered Session Management § § Easier to setup No SPOF Session reads are instantaneous Session writes generate Network I/O § Network I/O increases exponentially with increase in number of nodes § In very rare circumstances a request may get stale session data • User request reaches subsequent Clustered Session Management Load Balancer App. Server node faster than intra-node message • Intra-node communication fails § AKA Shared-nothing Cluster Creative Commons Sharealike Attributions Noncommercial 17

Load Balancer – Session Management • • Sticky Sessions with Central Session Store Sticky Sessions User 1 § Downtime does not cause loss of data § Session reads need not generate network I/O Sticky Sessions with Clustered Session Management § No specific advantages User 2 Load Balancer App. Server Creative Commons Sharealike Attributions Noncommercial App. Server 18

Load Balancer – Session Management • Recommendation § Use Clustered Session Management if you have – • Smaller Number of App Servers • Fewer Session writes § Use a Central Session Store elsewhere § Use sticky sessions only if you have to Creative Commons Sharealike Attributions Noncommercial 19

Load Balancer – Removing SPOF • • In a Load Balanced App Server Cluster the LB is an SPOF Setup LB in Active-Active or Active-Passive mode § Note: Active-Active nevertheless assumes that each LB is independently able to take up the load of the other § If one wants ZERO downtime, then Active-Active becomes truly cost beneficial only if multiple LBs (more than 3 to 4) are daisy chained as Active-Active forming an LB Cluster Active-Passive LB Users Load Balancer App. Server Active-Active LB Users Load Balancer App. Server Creative Commons Sharealike Attributions Noncommercial App. Server 20

Step 4 – Horizontal Scaling (App Server) • Load Balanced App Servers • Our deployment at the end of Step 4 Positives § Increases Availability and Scalability § No changes to App required § Easy setup DBServer • Negatives § Finite Scalability Creative Commons Sharealike Attributions Noncommercial 21

Step 5 – Vertical Partitioning (Hardware) • Load Balanced App Servers Introduction § Partitioning out the Storage function using a SAN • SAN config options § Refer to “Demystifying Storage” at http: //wiki. directi. com -> Dev University -> Presentations DBServer SAN • Positives § Allows “Scaling Up” the DB Server § Boosts Performance of DB Server • Negatives § Increases Cost Creative Commons Sharealike Attributions Noncommercial 22

Step 6 – Horizontal Scaling (DB) • § Increasing the number of DB nodes § Referred to as “Scaling out” the DB Server Load Balanced App Servers • DBServer Introduction DBServer Options § Shared nothing Cluster § Real Application Cluster (or Shared Storage Cluster) SAN Creative Commons Sharealike Attributions Noncommercial 23

Shared Nothing Cluster • • Each DB Server node has its own complete copy of the database Nothing is shared between the DB Server Nodes This is achieved through DB Replication at DB / Driver / App level or through a proxy Supported by most RDBMs natively or through 3 rd party software DBServer Database Note: Actual DB files maybe stored on a central SAN Creative Commons Sharealike Attributions Noncommercial 24

Replication Considerations • Master-Slave § Writes are sent to a single master which replicates the data to multiple slave nodes § Replication maybe cascaded § Simple setup § No conflict management required • Multi-Master § Writes can be sent to any of the multiple masters which replicate them to other masters and slaves § Conflict Management required § Deadlocks possible if same data is simultaneously modified at multiple places Creative Commons Sharealike Attributions Noncommercial 25

Replication Considerations • Asynchronous § § § • Guaranteed, but out-of-band replication from Master to Slave Master updates its own db and returns a response to client Replication from Master to Slave takes place asynchronously Faster response to a client Slave data is marginally behind the Master Requires modification to App to send critical reads and writes to master, and load balance all other reads Synchronous § Guaranteed, in-band replication from Master to Slave § Master updates its own db, and confirms all slaves have updated their db before returning a response to client § Slower response to a client § Slaves have the same data as the Master at all times § Requires modification to App to send writes to master and load balance all reads Creative Commons Sharealike Attributions Noncommercial 26

Replication Considerations • Replication at RDBMS level § Support may exists in RDBMS or through 3 rd party tool § Faster and more reliable § App must send writes to Master, reads to any db and critical reads to Master • Replication at Driver / DAO level § Driver / DAO layer ensures • writes are performed on all connected DBs • Reads are load balanced • Critical reads are sent to a Master § In most cases RDBMS agnostic § Slower and in some cases less reliable Creative Commons Sharealike Attributions Noncommercial 27

Real Application Cluster • • • All DB Servers in the cluster share a common storage area on a SAN All DB servers mount the same block device The filesystem must be a clustered file system (eg GFS / OFS) Currently only supported by Oracle Real Application Cluster Can be very expensive (licensing fees) DBServer Database Creative Commons Sharealike Attributions Noncommercial SAN 28

Recommendation • • • Try and choose a DB which natively supports Master. Slave replication Use Master-Slave Async replication Write your DAO layer to ensure § writes are sent to a single DB § reads are load balanced § Critical reads are sent to a master Load Balanced App Servers DBServer Writes & Critical Reads Other Reads Creative Commons Sharealike Attributions Noncommercial 29

Step 6 – Horizontal Scaling (DB) • Load Balanced App Servers DB DB Cluster • • Our architecture now looks like this Positives § As Web servers grow, Database nodes can be added § DB Server is no longer SPOF Negatives § Finite limit SAN Creative Commons Sharealike Attributions Noncommercial 30

Step 6 – Horizontal Scaling (DB) Reads Writes • DB 1 DB 2 • Shared nothing clusters have a finite scaling limit § Reads to Writes – 2: 1 § So 8 Reads => 4 writes § 2 DBs • Per db – 4 reads and 4 writes § 4 DBs • Per db – 2 reads and 4 writes § 8 DBs • Per db – 1 read and 4 writes At some point adding another node brings in negligible incremental benefit Creative Commons Sharealike Attributions Noncommercial 31

Step 7 – Vertical / Horizontal Partitioning (DB) • Load Balanced App Servers DB DB Cluster Introduction § Increasing the number of DB Clusters by dividing the data • Options § Vertical Partitioning - Dividing tables / columns § Horizontal Partitioning - Dividing by rows (value) SAN Creative Commons Sharealike Attributions Noncommercial 32

Vertical Partitioning (DB) • • Take a set of tables and move them onto another DB § Eg in a social network - the users table and the friends table can be on separate DB clusters Each DB Cluster has different tables Application code or DAO / Driver code or a proxy knows where a given table is and directs queries to the appropriate DB Can also be done at a column level by moving a set of columns into a separate table App Cluster DB Cluster 1 DB Cluster 2 Table 1 Table 2 Table 3 Table 4 Creative Commons Sharealike Attributions Noncommercial 33

Vertical Partitioning (DB) • Negatives § One cannot perform SQL joins or maintain referential integrity (referential integrity is as such overrated) § Finite Limit App Cluster DB Cluster 1 DB Cluster 2 Table 1 Table 2 Table 3 Table 4 Creative Commons Sharealike Attributions Noncommercial 34

Horizontal Partitioning (DB) • • Take a set of rows and move them onto another DB § Eg in a social network – each DB Cluster can contain all data for 1 million users Each DB Cluster has identical tables Application code or DAO / Driver code or a proxy knows where a given row is and directs queries to the appropriate DB Negatives App Cluster DB Cluster 1 DB Cluster 2 Table 1 Table 2 Table 3 Table 4 1 million users § SQL unions for search type queries must be performed within code Creative Commons Sharealike Attributions Noncommercial 35

Horizontal Partitioning (DB) • Techniques § FCFS • 1 st million users are stored on cluster 1 and the next on cluster 2 § Round Robin § Least Used (Balanced) • Each time a new user is added, a DB cluster with the least users is chosen § Hash based • A hashing function is used to determine the DB Cluster in which the user data should be inserted § Value Based • User ids 1 to 1 million stored in cluster 1 OR • all users with names starting from A-M on cluster 1 § Except for Hash and Value based all other techniques also require an independent lookup map – mapping user to Database Cluster § This map itself will be stored on a separate DB (which may further 36 need to be replicated) Creative Commons Sharealike Attributions Noncommercial

Step 7 – Vertical / Horizontal Partitioning (DB) • Load Balanced App Servers Lookup Map • • DB DB Cluster SAN • Our architecture now looks like this Positives § As App servers grow, Database Clusters can be added Note: This is not the same as table partitioning provided by the db (eg MSSQL) We may actually want to further segregate these into Sets, each serving a collection of users (refer next slide Creative Commons Sharealike Attributions Noncommercial 37

Step 8 – Separating Sets • Now we consider each deployment as a single Set serving a collection of users Global Lookup Map Global Redirector Load Balanced App Servers Lookup Map DB DB DB DB Cluster SAN SET 1 – 10 million users SET 2 – 10 million users Creative Commons Sharealike Attributions Noncommercial 38

Creating Sets • • • The goal behind creating sets is easier manageability Each Set is independent and handles transactions for a set of users Each Set is architecturally identical to the other Each Set contains the entire application with all its data structures Sets can even be deployed in separate datacenters Users may even be added to a Set that is closer to them in terms of network latency Creative Commons Sharealike Attributions Noncommercial 39

Step 8 – Horizontal Partitioning (Sets) • Global Redirector • App Servers Cluster DB Cluster SAN SET 1 • Our architecture now looks like this Positives § Infinite Scalability Negatives SET 2 § Aggregation of data across sets is complex § Users may need to be moved across Sets if sizing is improper § Global App settings and preferences need to be replicated across Sets Creative Commons Sharealike Attributions Noncommercial 40

Step 9 – Caching • • Add caches within App Server § § Object Cache Session Cache (especially if you are using a Central Session Store) API cache Page cache Software § Memcached § Teracotta (Java only) § Coherence (commercial expensive data grid by Oracle) Creative Commons Sharealike Attributions Noncommercial 41

Step 10 – HTTP Accelerator • • • If your app is a web app you should add an HTTP Accelerator or a Reverse Proxy A good HTTP Accelerator / Reverse proxy performs the following – § § § Redirect static content requests to a lighter HTTP server (lighttpd) Cache content based on rules (with granular invalidation support) Use Async NIO on the user side Maintain a limited pool of Keep-alive connections to the App Server Intelligent load balancing Solutions § Nginx (HTTP / IMAP) § Perlbal § Hardware accelerators plus Load Balancers Creative Commons Sharealike Attributions Noncommercial 42

Step 11 – Other cool stuff • • • CDNs IP Anycasting Async Nonblocking IO (for all Network Servers) If possible - Async Nonblocking IO for disk Incorporate multi-layer caching strategy where required § L 1 cache – in-process with App Server § L 2 cache – across network boundary § L 3 cache – on disk Grid computing § Java – Grid. Gain § Erlang – natively built in Creative Commons Sharealike Attributions Noncommercial 43

Platform Selection Considerations • • Programming Languages and Frameworks § Dynamic languages are slower than static languages § Compiled code runs faster than interpreted code -> use accelerators or pre-compilers § Frameworks that provide Dependency Injections, Reflection, Annotations have a marginal performance impact § ORMs hide DB querying which can in some cases result in poor query performance due to non-optimized querying RDBMS § § • My. SQL, MSSQL and Oracle support native replication Postgres supports replication through 3 rd party software (Slony) Oracle supports Real Application Clustering My. SQL uses locking and arbitration, while Postgres/Oracle use MVCC (MSSQL just recently introduced MVCC) Cache § Teracotta vs memcached vs Coherence Creative Commons Sharealike Attributions Noncommercial 44

Tips • • • All the techniques we learnt today can be applied in any order Try and incorporate Horizontal DB partitioning by value from the beginning into your design Loosely couple all modules Implement a REST-ful framework for easier caching Perform application sizing ongoingly to ensure optimal utilization of hardware Creative Commons Sharealike Attributions Noncommercial 45

Intelligent People. Uncommon Ideas. Questions? ? bhavin. t@directi. com http: //careers. directi. com Download slides: http: //wiki. directi. com 46