Denial-of-service attack

As explained in Distributed Machine Learning Patterns

A denial-of-service attack is an intentional or accidental activity that overwhelms a service with requests and prevents authorized access or normal operation.

A denial-of-service attack overwhelms a model-serving system with requests, interrupting authorized access or consuming shared computational resources needed by other services. The cause may be malicious, accidental, or a client misconfiguration. section 4.4.2

A practical defense uses rate limiting to control request pressure:

• Place incoming model-serving requests in a queue. section 4.4.2
• Process only the number of concurrent requests allowed by the current limit. section 4.4.2
• Continue with queued requests after active requests finish. section 4.4.2

Two-request concurrency limit

Four requests arrive, but a limit of two concurrent requests admits only two initially; the remaining requests wait until capacity becomes available. section 4.4.2

Figure 4.17 A flowchart of four model serving requests sent to the model serving system. However, only two are under the current rate limit, which allows a maximum of two concurrent model serving requests. Once the system has finished processing those two requests, it will proceed to the remaining two requests in the queue.

Visualizes how a rate-limiting queue admits requests in bounded groups and releases queued work as processing completes. section 4.4.2

How Distributed Machine Learning Patterns covers it

How deep it goes

Explains malicious and accidental causes, then gives a queue-based rate-limiting workflow with bounded concurrency.

Examples

Uses misconfigured clients, accidental load tests, unexpected traffic spikes, and four requests constrained to two concurrent operations.

What it assumes

Assumes a model-serving system that draws computational capacity from a shared resource pool.

What you can do after

Operate/configure request queues and concurrency limits to defend model-serving systems against excessive request load.

How it connects

Connects denial-of-service attacks to model serving, shared resource pools, event-driven processing, and rate limiting.

liveBookmanning.com

Read Distributed Machine Learning Patterns on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Secure APIsA denial-of-service attack is an attack intended to make a service or resource unavailable to legitimate users, including by exhausting its processing or storage resources.
• 

  liveBookmanning.com
  Acing the CCNA Exam: Volume 2 Advanced Networking and SecurityA denial-of-service attack is an attack that makes a system or service unavailable to legitimate users by exhausting or disrupting resources.
• 

  liveBookmanning.com
  Grokking Web Application SecurityA denial-of-service attack is an attack that makes a system or service unavailable or unusable by exhausting resources or overwhelming its processing capacity.

Compare how 4 books explain denial-of-service attack →

Related concepts

• rate limitaddressed by · A rate limit is a bound on the frequency or volume of operations permitted within a defined interval, scope, or concurrency limit.
• distributed denial of service attackincludes · A distributed denial-of-service attack uses traffic or requests from multiple sources to overwhelm a target and impair its availability or function.
• DHCP exhaustion attackincludes · A DHCP exhaustion attack is a denial-of-service technique that consumes a DHCP server's address pool so legitimate clients cannot obtain addresses.

This explainer was generated by Manning's knowledge graph from the text of Distributed Machine Learning Patterns. The section links point to the parts of the book it draws on.