1. CSE 490dp Resource Control Robert Grimm

2. Problems How to access resources? –Basic usage tracking How to measure resource consumption? –Accounting How to limit resource consumption? –Policy We only cover first two issues

3. Resource Access Issue –How to track usage? Four models –Don’t do it –Explicit creation and deletion –Automatic reclamation of unused resources –Leases

4. Don’t Do It Example: NFS [Ousterhout 90] –Most operations are idempotent –No connection state “Stateless” –Files persist “Stateful” by definition of a file system

5. Advantages Crash handling –No recovery on server –No reboots on client Simplicity –No connection handling –No recovery code

6. Problems Performance –No client-side caching Consistency –Clients don’t know about others modifying same file Semantics –Not implemented: locks –Not idempotent: mkdir

7. Explicit Creation and Deletion Example: Connections in Sprite FS [Ousterhout 90] Advantages –Consistency Server ensures only one writer per file –Performance Client-side caching –No synchronous writes to server –Files may be overwritten / deleted before they get sent to server

8. Problems Complexity –More to do on server Recovery –How to restore connections after failure? Performance –Explicit open and close operations Space overheads –Per-connection state

9. Pervasive Paradigm new / delete –C, C++ create / delete –Files –Tuples in T Spaces, one.world run / kill –Applications, servers

10. Automatic Reclamation Garbage collection Basic idea –Keep track of used resources –Automatically reclaim unused resources in background Examples –Locally for Java –Distributed for Java RMI

11. Discussion Advantage –No need to worry about deletion Problem –Complete and reliable usage information is hard to come by Network connections and nodes fail May depend on human factors –Hard to automate

12. Discussion Can we automate creation and deletion? –Cluster-based service Automatically run and kill services on individual nodes

13. Leases Basic idea –Limit resource access by time-out –Reclaim resource after time-out expires Operations –Acquire (open) –Renew –Cancel (close)

14. Examples Remote resources –Java RMI –Jini’s remote events, transactions File cache –V [Gray & Cheriton 89] Storage –JavaSpaces Resource access in general –one.world

15. Problem A resource can always vanish –Lease renewal is delayed E.g., message may be delayed, is lost –Lease expires

16. Tracking Usage Don’t do it Explicit creation and deletion Garbage collection –Automate deletion based on usage Leases –Automate deletion based on time-out

17. Resources Transient –Objects, connections Services –Applications User data –Files, tuples

18. Summary TransientServicesUser data Don’t do it?No ExplicitOK GC OK in intranets Fully automated! ? Leases Across the network ??

19. Break

20. Resource Accounting Problem –Traditionally, process represents Protection domain Resource principal –But, modern servers don’t follow this model [Banga et al., 99]

21. HTTP Server Implementations One process per connection Event-based server Multi-threaded server

22. One Process Per Connection

23. Event-Based Server

24. Multi-Threaded Server

25. Scheduling Entities and Activities Classical application –Single process –Mostly user space Modern applications –Considerable kernel processing –Multiple processes –Multiple threads

26. Classical Application

27. Considerable Kernel Processing

28. Multiple Processes

29. Multiple Threads

30. How to Account for Resources? Introduce new abstraction –Resource container –Tracks usage information CPU time Memory Networking bandwidth … –Orthogonal to scheduling entities Threads, processes have resource binding

31. Scheduler Binding Problem –How to schedule threads that are bound to several resource containers? E.g., event-based server Solution –Scheduler binding Set of all resource containers

32. Operations Create Set parent –Hierarchy of containers Release –Reference counted Share Access attributes Access usage information

33. Operations Bind thread to a container Reset scheduler binding Bind socket, file to a container

34. Usage One resource container per logical job –Create new container when accepting connection –Change thread’s container as necessary Example policy –Different classes of clients, depending on source IP address Low priority High priority

35. Performance Growing number of low priority clients

36. Performance Growing number of concurrent CGI requests

37. Performance Growing SYN-flood rate

38. Summary Resource accounting is orthogonal to –Protection –Scheduling Need explicit abstraction –Resource containers

39. Issues How to express and enforce policies? How to make resource containers scale across the network?