Igor Korotach
Head of FinTech at Quantum
Choosing the right tool for the job!
Written by: Igor Korotach
Which types of IT projects do you know?
By type of work
- Web Development
- Software Development
- Mobile App Development
- Network Configuration
- Software Implementation
- Hardware Installation
- Database Management
- IT Emergency Recovery
By business domain
- Banking, Finance Services and Insurance
- Manufacturing
- E-commerce
- Online Services and marketing
- Logistics and Supply Chain Management
- Telecommunication
- Pure IT industry
By client's attitude...
Hype train
Ol' reliable
"It's a certified IT technology, that..."
Should have been done yesterday
Gotta go fast!
Types of architecture
- Monolith with layers (N-layered)
- Service Oriented Architecture
- Event Driven Architecture
- Pipes and Filters
- Cloud Native Architecture
Tech choice diagram
Performance, Reliability, Ease of Use
Performance criteria
- System metrics (CPU, RAM, GPU, etc.)
- Request metrics (Request per second, Average response time, etc.)
- Network metrics (Latency, bandwidth, etc.)
- Scalability metrics (scalability factor, scalability curve, etc.)
Reliability criteria
- Error metrics (Rate of errors, Criticality of errors, etc.)
- Availability metrics (Uptime, etc.)
- Security metrics (Security incidents, CVE, Issues, etc.)
- Licensing policy
- Production readiness (Readiness of the docs, Time to resolve issues, Customer/Community Support, Feature richness)
- Open sourceness & developer support
- Resource availability (amount of engineers who know that tech on the market/in the company)
Ease of Use criteria
- Documentation quality
- Getting started section
- Productivity guidelines
- Feature richness & explanation
- Boilerplate code amount
- Speed of development for the tool
Rules of choosing a technology
- Understand the surrounding restrictions & possibilities
- Find the option you like
- Find an alternative traditional solution
- Test the claims given on your options
- Find the pros of your choice
- Find the cons of your choice
Example: Redpanda
Claimed
Independently tested
Redpanda Performance
Kafka taken for comparison
https://jack-vanlightly.com/blog/2023/5/15/kafka-vs-redpanda-performance-do-the-claims-add-up
Redpanda
Kafka
Redpanda Reliability
Kafka taken for comparison
Redpanda
Kafka
Redpanda Ease of Use
Kafka taken for comparison
Pros
Cons
Redpanda Pros & Cons
- Performance - Low throughput is where Redpanda has its most robust end-to-end latency results, then as you increase throughput, the specifics of the workload can have a disproportionate impact on performance.
- Ease of use - Client libraries are compatible with Kafka, Getting started is easier
- Licensing - Business Source License
- Performance - Several common production configurations, such as high producer count or running for more than 24 hours can cause severe reductions in performance.
- Reliability - Limited adoption
- Reliability - features are limited (e..g. Stream Processing)
Personal tips on choosing the right tool
- When looking at new piece of tech start with R&D
- When starting a project a general purpose tool is better than a niche tailored one
- Don't forget 'Make it work -> make it fast'
- Good abstractions + bad tech >= bad abstractions + good tech
- Don't get into 'it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail'
- Sometimes you have to stop sharpening the saw, and just start cutting shit
- Be skeptical but curious about new tools
Everything comes with a tradeoff
Thanks for your attention. You've been awesome!
Questions?
- Presentation link: https://slides.com/emulebest/choosing-the-right-tool
Choosing the right tool
By Igor Korotach
