Disco Running Commodity Operating Systems on Scalable Multiprocessors

Disco: Running Commodity Operating Systems on Scalable Multiprocessors Edouard Bugnion, Scott Devine, Mendel Rosenblum, Stanford University, 1997 Presented by Divya Parekh 1

Outline n n Virtualization Disco description Disco performance Discussion 2

Virtualization n “a technique for hiding the physical characteristics of computing resources from the way in which other systems, applications, or end users interact with those resources. This includes making a single physical resource appear to function as multiple logical resources; or it can include making multiple physical resources appear as a single logical resource” 3

Old idea from the 1960 s n IBM VM/370 – A VMM for IBM mainframe Multiple OS environments on expensive hardware § Desirable when few machine around § n Popular research idea in 1960 s and 1970 s Entire conferences on virtual machine monitors § Hardware/VMM/OS designed together § § Interest died out in the 1980 s and 1990 s Hardware got more cheaper § Operating systems got more powerful (e. g. multi-user) § 4

A Return to Virtual Machines n Disco: Stanford research project (SOSP ’ 97) n n n Commercial virtual machines for x 86 architecture n n n VMware Workstation (now EMC) (1999 -) Connectix Virtual. PC (now Microsoft) Research virtual machines for x 86 architecture n n n Run commodity OSes on scalable multiprocessors Focus on high-end: NUMA, MIPS, IRIX Xen (SOSP ’ 03) plex 86 OS-level virtualization n Free. BSD Jails, User-mode-linux, UMLinux 5

Overview n Virtual Machine n n A fully protected and isolated copy of the underlying physical machine’s hardware. (definition by IBM)” Virtual Machine Monitor n n A thin layer of software that's between the hardware and the Operating system, virtualizing and managing all hardware resources. Also known as “Hypervisor” 6

Classification of Virtual Machines 7

Classification of Virtual Machines n Type I n n VMM is implemented directly on the physical hardware. VMM performs the scheduling and allocation of the system’s resources. IBM VM/370, Disco, VMware’s ESX Server, Xen Type II n n n VMMs are built completely on top of a host OS. The host OS provides resource allocation and standard execution environment to each “guest OS. ” User-mode Linux (UML), UMLinux 8

Non-Virtualizable Architectures n According to Popek and Goldberg, ” an architecture is virtualizable if the set of sensitive instructions is a subset of the set of privileged instructions. ” n x 86 n n Several instructions can read system state in register CPL 3 without trapping MIPS n KSEG 0 bypasses TLB, reads physical memory directly 9

Type I contd. . n Hardware Support for Virtualization Figure: The hardware support approach to x 86 Virtualization E. g. Intel Vanderpool/VT and AMD-V/SVM 10

Type I contd. . n Full Virtualization Figure : The binary translation approach to x 86 Virtualization E. g. VMware ESX server 11

Type I contd. . n Paravirtualization Figure: The Paravirtualization approach to x 86 Virtualization E. g. Xen 12

Type II n Hosted VM Architecture E. g. VMware Workstation, Connectix Virtual. PC 13

Disco : VMM Prototype n Goals n Extend modern OS to run efficiently on shared memory multiprocessors without large changes to the OS. n A VMM built to run multiple copies of Silicon Graphics IRIX operating system on a Stanford Flash shared memory multiprocessor. 14

Problem Description n Multiprocessor in the market (1990 s) n Innovative Hardware faster than System Software n Customized OS are late, incompatible, and possibly bug Commodity OS not suited for multiprocessors n Do not scale cause of lock contention, memory architecture n Do not isolate/contain faults n More Processors More failures 15

Solution to the problems n n Resource-intensive Modification of OS (hard and time consuming, increase in size, etc) Make a Virtual Machine Monitor (software) between OS and Hardware to resolve the problem 16

Two opposite Way for System Software n n Address these challenges in the operating system: OS -Intensive n Hive , Hurricane, Cellular-IRIX, etc n innovative, single system image n But large effort. Hard-partition machine into independent failure units: OS-light n Sun Enterprise 10000 machine n Partial single system image n Cannot dynamically adapt the partitioning 17

Return to Virtual Machine Monitors n n n One Compromise Way between OS-intensive & OSlight – VMM Virtual machine monitors, in combination with commodity and specialized operating systems, form a flexible system software solution for these machines Disco was introduced to allow trading off between the costs of performance and development cost. 18

Architecture of Disco 19

Advantages of this approach n n n Scalability Flexibility Hide NUMA effect Fault Containment Compatibility with legacy applications 20

Challenges Facing Virtual Machines n n n Overheads n Trap and emulate privileged instructions of guest OS n Access to I/O devices n Replication of memory in each VM Resource Management n Lack of information to make good policy decisions Communication and Sharing n Stand alone VM’s cannot communicate 21

Disco’s Interface n Processors n n n MIPS R 10000 processor Emulates all instructions, the MMU, trap architecture Extension to support common processor operations n n Physical memory n n Enabling/disabling interrupts, accessing privileged registers Contiguous, starting at address 0 I/O devices n n n Virtualize devices like I/O, disks, n/w interface exclusive to VM Physical devices multiplexed by Disco Special abstractions for SCSI disks and network interfaces n n Virtual disks for VMs Virtual subnet across all virtual machines 22

Disco Implementation n n Multi threaded shared memory program Attention to NUMA memory placement, cache aware data structures and IPC patterns Code segment of DISCO copied to each flash processor – data locality Communicate using shared memory 23

Virtual CPUs n Direct Execution n n Challenges n n n execution of virtual CPU on real CPU Sets the real machine’s registers to the virtual CPU’s Jumps to the current PC of the virtual CPU, Direct execution on the real CPU Detection and fast emulation of operations that cannot be safely exported to the virtual machine privileged instructions such as TLB modification and Direct access to physical memory and I/O devices. Maintains data structure for each virtual CPU for trap emulation Scheduler multiplexes virtual CPU on real processor 24

Virtual Physical Memory n n n Address translation & maintains a physical-tomachine address (40 bit) mapping. Virtual machines use physical addresses Software reloaded translation-lookaside buffer (TLB) of the MIPS processor Maintains pmap data structure for each VM – contains one entry for each physical to virtual mapping pmap also has a back pointer to its virtual address to help invalidate mappings in the TLB 25

Contd. . n n Kernel mode references on MIPS processors access memory and I/O directly - need to re-link OS code and data to a mapped address space MIPS tags each TLB entry with Address space identifiers (ASID) ASIDs are not virtualized - TLB need to be flushed on VM context switches Increased TLB misses in workloads n n n Additional Operating system references VM context switches TLB misses expensive - create 2 nd level software TLB. Idea similar to cache? 26

NUMA Memory management n n Cache misses should be satisfied from local memory (fast) rather than remote memory (slow) Dynamic Page Migration and Replication n n Pages frequently accessed by one node are migrated Read-shared pages are replicated among all nodes Write-shared are not moved, since maintaining consistency requires remote access anyway Migration and replacement policy is driven by cache-misscounting facility provided by the FLASH hardware 27

Transparent Page Replication 1. Two different virtual processors of the same virtual machine logically read-share the same physical page, but each virtual processor accesses a local copy. 2. memmap tracks which virtual page references each physical page. Used during TLB shootdown 28

Disco Memory Management 29

Virtual I/O Devices n n Disco intercepts all device accesses from the virtual machine and forwards them to the physical devices Special device drivers are added to the guest OS Disco device provide monitor call interface to pass all the arguments in single trap Single VM accessing a device does not require virtualizing the I/O – only needs to assure exclusivity 30

Copy-on-write Disks n n n Intercept DMA requests to translate the physical addresses into machine addresses. Maps machine page as read only to destination address page of DMA Sharing machine memory Attempts to modify a shared page will result in a copy -on-write fault handled internally by the monitor. n n n Logs are maintained for each VM Modification made in main memory Non-persistent disks are copy on write shared n n E. g. Kernel text and buffer cache E. g. File systems root disks 31

Transparent Sharing of Pages Creates a global buffer cache shared across VM's and reduces memory foot print of the system 32

Virtual Network Interface n n n Virtual subnet and network interface use copy on write mapping to share the read only pages Persistent disks can be accessed using standard system protocol NFS Provides a global buffer cache that is transparently shared by independent VMs 33

Transparent sharing of pages over NFS 1. The monitor’s networking device remaps the data page from the source’s machine address space to the destination’s. 2. The monitor remaps the data page from the driver’s mbuf to the clients buffer cache. 34

Modifications to the IRIX 5. 3 OS n Minor changes to kernel code and data segment – specific to MIPS n n n Relocate the unmapped segment of the virtual machine into the mapped supervisor segment of the processor– Kernel relocation Disco drivers are same as original device drivers of IRIX Patched HAL to use memory loads/stores instead of privileged instructions 35

Modifications to the IRIX 5. 3 OS n n Added code to HAL to pass hints to monitor for resource management New Monitor calls to MMU to request zeroed page, unused memory reclamation Changed mbuf management to be pagealigned Changed bcopy to use remap (with copy-onwrite) 36

SPLASHOS: A specialized OS n n n Thin specialized library OS, supported directly by Disco No need for virtual memory subsystem since they share address space Used for the parallel scientific applications that can span the entire machine 37

Disco: Performance n Experimental Setup n n n Disco targets the FLASH machine not available that time Used Sim. OS, a machine simulator that models the hardware of MIPS-based multiprocessors for the Disco monitor. Simulator was too slow to allow long work loads to be studied 38

Disco: Performance n Workloads 39

Disco: Performance n Execution Overhead Pmake overhead due to I/O virtualization, others due to TLB mapping Reduction of kernel time On average virtualization overhead of 3% to 16% 40

Disco: Performance n Memory Overheads V: Pmake memory used if there is no sharing M: Pmake memory used if there is sharing 41

Disco: Performance n Scalability Partitioning of problem into different VM’s increases scalability. Kernel synchronization time becomes smaller. 42

Disco: Performance n Dynamic Page Migration and replication 43

Conclusion n Disco VMM hides NUMA-ness from non. NUMA aware OS Disco VMM is low(er) effort Moderate overhead due to virtualization 44

Discussion n Was Disco- VMM done rightly? n Virtual Physical Memory on architectures other than MIPS n n n MIPS TLB is software managed Not sure of how well other OS perform on Disco since IRIX was designed for MIPS Not sure how HIVE, Hurricane performs comparatively Performance of long workloads on the system Performance of heterogeneous VMs e. g. Pmake case 45

Discussion Are VMM Microkernels done right? 46