???? 2020/04/17
????课件 ch17 & 18
graph database

---

CH17

Transaction

Transaction concept

ACID Properties

Transaction State

Schedules

Schedule 1

Schedule 2

Schedule 3

Schedule 4

Serializability

Simplified view of transactions

Conflicting Instructions

Conflict Serializability

conflict serializable example

NOT conflict serializable example

View Serializability

(prof skip)

Testing for Serializability

vertices 顶点/交点
precedence 优先

Test for Conflict Serializability

Test for View Serializability

(skip)

Recoverable Schedules

Cascading Rollbacks

Concurrency Control

Weak Levels of Consistency

Levels of Consistency in SQL-92

---

CH18

Concurrency Control

Lock-Based Protocols

Protocols 礼仪/规章制度

You can have multiple share locks at the same time, but you can ONLY have ONE exclusive lock.

Schedule With Lock Grants

Deadlock

The Two-Phase Locking Protocol

Deadlock

Deadlock Handling

Deadlock prevention protocols ensure that the system will never enter into a deadlock state. Some prevention strategies:

• Require that each transaction locks all its data items before it begins execution (pre-declaration).
• Impose partial ordering of all data items and require that a transaction can lock data items only in the order specified by the partial order (graph-based protocol).

Deadlock Detection

Deadlock Recovery

Multiple Granularity

Example of Granularity Hierarchy

Intention Lock Modes

Insert/Delete Operations and Predicate Reads

Phantom Phenomenon

Phantom Phenomenon Example

Handling Phantoms

Index Locking To Prevent Phantoms

Timestamp-Based Protocols

Validation-Based Protocols

NoSQL Databases

Introduction

One Taxonomy

Taxonomy 分类

CAP Theorem

CAP Theorem

Consistency Models

• tradeoff in the CAP theorem
  So in a replicated database, I can get consistancy through locking, but I can’t function of this partition caz I can’t lock enough copies. And so that’s where the partition tolerance comes in. For availability, I need multiple copies. For consistency, I need to lock all the copies, which means I can’t tolerate partitions. So I can have consistency and availability, but not partition tolerance. If I have partition tolerance, by definition I can’t lock all the copies, which means I can’t have consistency. I can get partition tolerance and consistency, but not availabilty.