Do you have a high-value, high throughput application running on AWS? Are you moving part or all of your infrastructure to AWS? Do you have a high-transaction workload that is only expected to grow as your company grows? Choosing the right database for your move to AWS can make you a hero or a goat. Be a hero!
Databases are the mission-critical lifeline of most businesses. For years MySQL has been the easy choice -- but the popularity of the cloud and new products like Aurora, RDS MySQL and ClustrixDB have given customers choices and options that can help them work smarter and more efficiently.
Enterprise Strategy Group (ESG) presents their findings from a recent performance benchmark test configured for high-transaction, low-latency workloads running on AWS.
In this webinar, you will learn:
How high-transaction, high-value database workloads perform when run on three popular databases solutions running on AWS.
How key metrics like transactions per second (tps) and database response time (latency) can affect performance and customer satisfaction.
How the ability to scale both database reads and writes is the key to unlocking performance on AWS
2. About Enterprise Strategy Group
• ESG is an IT analyst, research, and strategy company.
• Our firm was founded in 1999 with headquarters in Milford, MA / an analyst and client relations presence in
Silicon Valley, CA.
• ESG conducts research with and for IT vendors, IT professionals, business professionals, and channel partners.
• We maintain ongoing analyst coverage in cloud computing, networking, storage, data protection, cybersecurity,
data management and analytics, application development and deployment, enterprise mobility, and channels.
• Capabilities include: Analyst services, market research, technical and economic validation, consulting, and
custom content.
• Mike Leone is a Senior ESG Lab Analyst with coverage across the entire IT industry.
• Provide in-depth testing and analysis of IT technology and products, using methods that simulate or recreate real-world environments.
• Extensive background in performance modeling, testing, and analysis.
9. ClustrixDB: Scale-Out, Fault-tolerant, MySQL-Compatible
ClustrixDB Overview
9
ClustrixDB
ACID Complaint
Transactions & Joins
Optimized for OLTP
Built-In Fault Tolerance
Flex-Up and Flex-Down
Minimal DB Admin
Also runs great in
the Data Center
Built to run
in the Cloud
10. ClustrixDB Technical Overview
Fully Distributed & Consistent Cluster
• Fully Consistent, and ACID-compliant database
– Supports transactions
– Support joins
– Optimized for OLTP
– But also supports reporting SQL
• All nodes are equal (no “special” node)
• All servers are read/write
• All servers accept client connections
• Tables & Indexes distributed across all nodes
– Fully automatic distribution, re-balancing
& re-protection
ClustrixDB Overview10
PrivateNetwork
ClustrixDB on commodity/cloud servers
HW or SW Load
Balancer
SQL-based
Applications
High Concurrency
Custom:
PHP, Java, Ruby, etc
Packaged:
Magento, etc
12. S1
S2
S3
S3
S4
S4
S5
Database Capacity
• Easy and simple Flex Up (and Flex Down)
– Single minimal interruption of service
• All servers handle writes + reads
• Data is automatically rebalanced
across the cluster
ClustrixDB Overview12
S1
ClustrixDB
S2
S5
13. S1
S2
S3
S3
S4
S4
S5
Built-in Fault Tolerance
• Server node goes down…
– Data is automatically rebalanced across
the remaining nodes
• Simply Add new Node
– System automatically re-protects
– Data is automatically redistributed
ClustrixDB Overview13
S1
ClustrixDB
S2
S5
14. ClustrixDB Rebalancer: Making the Complex, Simple
Q: How do you ensure data stays well distributed in a clustered environment?
A: You let the Rebalancer handle it!
The Rebalancer automatically:
• Initial Data: Distributes the data into even slices across nodes
• Data Growth: Splits large slices into smaller slices
• Flex-Up/Flex-Down: Moves slices to leverage new nodes and/or evacuate nodes
• Failed Nodes: Re-protects slices to ensure proper replicas exist
• Skewed Data: Re-distributes the data to even out across nodes
• Hotness Detection: Finds hot slices and balances then across nodes
…while the DB stays open for businessPatent 8,543,538
Patent 8,554,726
23. Making a Representative Benchmark
PROPRIETARY AND CONFIDENTIAL23
Why did we choose the workload that we did?
• Other venders recently have published performance results with Sysbench.
• Things that we think matter:
– Read/Write Mix Matters: Real workloads are a mixture of reads and writes.
– Data Size Matters: OLTP databases, while not “big data”, still have record counts of 10
million or more.
– Latency Matters: OLTP workloads demand low latency.
• The ClustrixDB Sysbench Performance Benchmark
– Mixed workload: 90% read and 10% writes
– Size: 40m records / 20 GB file
– Latency: 20ms is the upper range of “acceptable” latency (>3X the initial latency).
• Methodology: Performance curves show a more complete story.
– Reveals the “capacity point of the machine”
• The “knee of the curve” where throughput gains succumb to growth in latency.
25. ClustrixDB Key Differentiators
• MySQL Compatible, Scale-out ACID RDBMS
• Flex-Up and Flex-Down
– Flex Licensing allows paying for only the instances you need
– Add or remove multiple nodes at a time
• Massively Scaling Writes
– ClustrixDB is the only cloud RDBMS that can massively scale
writes in addition to reads
– E-commerce, AdTech, IoT are very write-intensive. Clustrix
has multiple satisfied customers in these sectors
• Built for Cloud (any cloud) and Datacenter
– One RDBMS wherever the application is deployed
PROPRIETARY AND CONFIDENTIAL25