snowbird: interactive resource-intensive applications made easy

Snowbird:Interactive Resource-Intensive

Applications Made Easy

H. Andrés Lagar-Cavilla*

Niraj Tolia‡, Eyal de Lara*,

M. Satyanarayanan‡ & Dave O’Hallaron‡

*U of Toronto, ‡CMU, September 2007

Bimodal Applications

• Interactive– Cognitive phase

• Resource-Intensive– Crunch phase

• Digital Animation• Scientific Computing• Engineering Design• Bio/Pharma• Video Editing• ….

Maya

(digital animation)

Different Phases, Different Needs

• Crunch (computation)– Short completion time

– Computing resources

• Cognitive (interaction)– Crisp interactive performance

– User attention

What To Do?Execution alternatives

– Thick Clients (Desktop PCs) Have excellent interactive performance

• Can be resource constrained

– Thin Clients (VNC, Remote Desktop) Allow use of remote resources (CPU, Data)

• Latency and jitter impact interactive performance

– Custom Applications Pipeline: placement over different nodes

• Requires significant developer resources

What To Do? A Systems Solution

With the best of both worlds Have excellent interactive performance

Allow use of remote resources (CPU, Data)

Internet

ImprovedCompute Power

Compute Cluster

Snowbird: Agent Abstraction

Applications encapsulated within an agent– Agent: processes, libraries, IPC, OS, config data…– Migration: performance goals achieved by morphing

• a thin client for resource intensive crunch phases• a thick client for highly interactive cognitive phases

ImprovedData Access

User’s Desktop Data Repository

ImprovedInteractivity

Agent

Bimodal Applications Made Easy

• Develop apps as monolithic blocks– Don’t worry about what executes where– Agent is migrated to satisfy each phase

• Seamless and transparent behavior

• Legacy support– Different OS’s (and versions/features)– Different languages (Fortran!)– No need for recompilation, relinking, etc…– Closed source apps just work

Design Criteria• VM-based migration

– x86 interface most widely deployed– Transparently support OS, lang, etc…

• Internet scale– WAN migration– Long fat pipes

• 50 Mbp/s…, 50-200 ms RTT

• Server-less design (P2P)– All hosts symmetric– Can execute anywhere

Implementation

Implementation

• VMM: suspend, resume, live migration – Xen 3.0.1

• Interaction-aware migration manager– Transparent

• Support for 3D acceleration in VMs– Vital for crisp interaction

• Virtual disk that maximizes locality– WAN area migration

Agent Profiles: Migration Manager

• FSM that models an agent’s behavior• Provided by expert users, admins, or developers

– Default system-wide profile available– Only deployment additional effort

CPUIntensive

cycles.org

NetIntensive

data.edu

InteractionIntensiveusr.home

Net > 4 Mbit/s

CPU > 95%

FPS < 20 &&

Input > 15

Internet

ImprovedCompute Power

Compute Cluster

ImprovedData Access

User’s Desktop Data Repository

ImprovedInteractivity

Agent

Interaction-aware

• Novelty in our approach

• Measure the quality of the interactive response

• We use frames per seconds– More in sync with operations in bimodal apps– Stretch object, rotate, zoom, etc…– Pure latency not enough

• Ample space for future work

Frames Per Second

• Non work-conserving (VNC): FPS = 2/latency

• Local: FPS = n/latency

• Work-conserving (X): FPS = n/(latency*k)

VMGL: 3D Acceleration in VMs

• OpenGL virtualization

• Hardware specs closed, unavailable

• Focus instead on software standards– OpenGL -> cross-platform– Direct3D -> MS-only

• Intercept GL calls and forward them to the host. Proprietary driver renders there.

• More details: VEE 2007

Open GL for X11 Apps

VMGL for X11 Apps

Admin VM

LocalChunk Store

WANDisk Manager

Kernel Module

WANDisk: Virtual Disk

foo.toronto.edu

Admin VM

LocalChunk Store

WANDisk Manager

Kernel Module

bar.cmu.edu

ChunkTable

ChunkMisses

Agent

Block DevOS

Application

baz.europe.org

Admin VM

LocalChunk Store

WANDisk Manager

Kernel Module

Why Another Storage System?• Exploits Snowbird characteristics

• P2P model– No server interposition– Single-writer: simple metadata, no locks

• Minimizes WAN talk– Locality: persistent replicas– Differential transfers: rsync– On-demand fetching

Evaluation

Benchmarks

Maya (closed source)ADF (closed source)

QuakeViz Kmenc15

• Broad set of domains

– Animation, Scientific Computing, Video Editing, Bio

• Closed and open source

• Straightforward installation

• Able to use generic profile on all four

• Partitioned mode for comparison

– Adf & Maya

Methodology• Crunch + cognitive benchmarks• Performing “crunch” experiments is easy• Replaying long interactive traces is not

– Can’t expect a user to do it– Must be able to compare results

• VNC-Redux: record and replay interactive user sessions– Record input and screen state – This is matched during replay for accuracy

Experimental Setup

• Thick: No virtualization, on User Desktop– User Desktop: UP with graphics acceleration

• Thin: No virtualization, on Compute Server– Compute Server: 4-way SMP– 100 Mbit/s WAN, RTT: 33, 66, and 100 ms

• Partitioned– App-specific developer-brewed: Maya & ADF

• Snowbird– Agents are initially launched on User Desktop

Results: Crunch Phase

45

45

67

21

96

10

7

12

5

434

7

57

74

23

48

58

76

24

49

58

78

24

45

0

20

40

60

80

100

120

140

Maya QuakeViz ADF Kmenc15

Min

ute

s

ThinThickSnowbird 33Snowbird 66Snowbird 100Partitioned

Snowbird’s crunch performance– Much better than thick – Comparable to thin/partitioned

Ouch!

• Better interactivity than thin clients• Is Snowbird any worse than a thick client?• > 20 FPS is ok, < 8 FPS is unusable

Results: Cognitive Phase

Snowbird Limitations

• Parallelism up to SMP level– What about cluster-scale?

• SSE, 3DNow!– i.e. x86 is not that uniform

• Overlapping phases– Hysteresis, priority in migration manager

• Very short phases– Cost/benefit analysis

Take Home Messages

• Bimodal applications – What they are and why they matter

• Thin clients are not almighty– There is no replacement for local interaction

• Best of both worlds: thick and thin clients– Necessary in an Internet world with remote computing

resources

• VM-based app migration is feasible– And with many advantages

• Future trends align well…

Futurism

• More bandwidth: cheaper VM migration

• What about latency?– The earth is not shrinking– Speed of light is not increasing– More routers, overlays, firewalls– Europe: ~100ms (60ms speed-of-light)– Asia: ~200ms (90ms speed-of-light)

• Insurmountable obstacle for thin clients

Our Current Interests

• VM support for large parallel tasks– Relevant to commodity computing– Migration, the cloud, etc…

• How to measure interactive performance– Thin clients, desktop consolidation, VMs– Unknown effects for modern (3D-heavy) GUIs

Questions?

Thanks More details in Middleware 2007

H. Andrés Lagar-Cavillaandreslc@cs.toronto.edu

Niraj Tolia, Eyal de Lara, Satya & Dave O’HallaronU of Toronto, CMU

Backup

Frames Per Second

• Non work-conserving: same latency, less frames

• Local: FPS = frames/latency

• Work-conserving: same frames, more latency

Applicability of Snowbird

• Morphing time: our implementation• Speedup: application & resources• C: Crunch phase time locally

Keywords Of This Presentation

• Thin clients– Remote execution

• Interactive Performance– Thick clients

• Virtual machine migration– Application migration

• Bimodal applications – What they are and why they matter

What We Need

• System support for bimodal applications

• Combine best of both worlds– Thick client, local execution

• Interaction

– Thin client, remote execution• Computation

• Make development easy

Results: Crunch Phase

0

20

40

60

80

100

120

140

Maya QuakeViz ADF Kmenc15

Application

Tim

e (

Min

ute

s)

ThinThickSnowbird 33Snowbird 66Snowbird 100

Snowbird’s crunch performance– Much better than thick – Comparable to thin/partitioned