Random Notes
  • Introduction
  • Reading list
  • Theory
    • Index
      • Impossibility of Distributed Consensus with One Faulty Process
      • Time, Clocks, and the Ordering of Events in a Distributed System
      • Using Reasoning About Knowledge to analyze Distributed Systems
      • CAP Twelve Years Later: How the “Rules” Have Changed
      • A Note on Distributed Computing
  • Operating System
    • Index
  • Storage
    • Index
      • Tachyon: Reliable, Memory Speed Storage for Cluster Computing Frameworks
      • Exploiting Commutativity For Practical Fast Replication
      • Don’t Settle for Eventual: Scalable Causal Consistency for Wide-Area Storage with COPS
      • Building Consistent Transactions with Inconsistent Replication
      • Managing Update Conflicts in Bayou, a Weakly Connected Replicated Storage System
      • Spanner: Google's Globally-Distributed Database
      • Bigtable: A Distributed Storage System for Structured Data
      • The Google File System
      • Dynamo: Amazon’s Highly Available Key-value Store
      • Chord: A Scalable Peer-to-peer Lookup Service for Internet Applications
      • Replicated Data Consistency Explained Through Baseball
      • Session Guarantees for Weakly Consistent Replicated Data
      • Flat Datacenter Storage
      • Small Cache, Big Effect: Provable Load Balancing forRandomly Partitioned Cluster Services
      • DistCache: provable load balancing for large-scale storage systems with distributed caching
      • Short Summaries
  • Coordination
    • Index
      • Logical Physical Clocks and Consistent Snapshots in Globally Distributed Databases
      • Paxos made simple
      • ZooKeeper: Wait-free coordination for Internet-scale systems
      • Just Say NO to Paxos Overhead: Replacing Consensus with Network Ordering
      • Keeping CALM: When Distributed Consistency is Easy
      • In Search of an Understandable Consensus Algorithm
      • A comprehensive study of Convergent and Commutative Replicated Data Types
  • Fault Tolerance
    • Index
      • The Mystery Machine: End-to-end Performance Analysis of Large-scale Internet Services
      • Gray Failure: The Achilles’ Heel of Cloud-Scale Systems
      • Capturing and Enhancing In Situ System Observability for Failure Detection
      • Check before You Change: Preventing Correlated Failures in Service Updates
      • Efficient Scalable Thread-Safety-Violation Detection
      • REPT: Reverse Debugging of Failures in Deployed Software
      • Redundancy Does Not Imply Fault Tolerance
      • Fixed It For You:Protocol Repair Using Lineage Graphs
      • The Good, the Bad, and the Differences: Better Network Diagnostics with Differential Provenance
      • Lineage-driven Fault Injection
      • Short Summaries
  • Cloud Computing
    • Index
      • Improving MapReduce Performance in Heterogeneous Environments
      • CLARINET: WAN-Aware Optimization for Analytics Queries
      • MapReduce: Simplified Data Processing on Large Clusters
      • Dryad: Distributed Data-Parallel Programs from Sequential Building Blocks
      • Resource Management
      • Apache Hadoop YARN: Yet Another Resource Negotiator
      • Mesos: A Platform for Fine-Grained Resource Sharing in the Data Center
      • Dominant Resource Fairness: Fair Allocation of Multiple Resource Types
      • Large-scale cluster management at Google with Borg
      • MapReduce Online
      • Delay Scheduling: A Simple Technique for Achieving Locality and Fairness in Cluster Scheduling
      • Reining in the Outliers in Map-Reduce Clusters using Mantri
      • Effective Straggler Mitigation: Attack of the Clones
      • Resilient Distributed Datasets: A Fault-Tolerant Abstraction for In-Memory Cluster Computing
      • Discretized Streams: Fault-Tolerant Streaming Computation at Scale
      • Sparrow: Distributed, Low Latency Scheduling
      • Making Sense of Performance in Data Analytics Framework
      • Monotasks: Architecting for Performance Clarity in Data Analytics Frameworks
      • Drizzle: Fast and Adaptable Stream Processing at Scale
      • Naiad: A Timely Dataflow System
      • The Dataflow Model:A Practical Approach to Balancing Correctness, Latency, and Cost in Massive-Scale
      • Interruptible Tasks:Treating Memory Pressure AsInterrupts for Highly Scalable Data-Parallel Program
      • PACMan: Coordinated Memory Caching for Parallel Jobs
      • Multi-Resource Packing for Cluster Schedulers
      • Other interesting papers
  • Systems for ML
    • Index
      • A Berkeley View of Systems Challenges for AI
      • Tiresias: A GPU Cluster Managerfor Distributed Deep Learning
      • Gandiva: Introspective Cluster Scheduling for Deep Learning
      • Workshop papers
      • Hidden Technical Debt in Machine Learning Systems
      • Inference Systems
      • Parameter Servers and AllReduce
      • Federated Learning at Scale - Part I
      • Federated Learning at Scale - Part II
      • Learning From Non-IID data
      • Ray: A Distributed Framework for Emerging AI Applications
      • PipeDream: Generalized Pipeline Parallelism for DNN Training
      • DeepXplore: Automated Whitebox Testingof Deep Learning Systems
      • Distributed Machine Learning Misc.
  • ML for Systems
    • Index
      • Short Summaries
  • Machine Learning
    • Index
      • Deep Learning with Differential Privacy
      • Accelerating Deep Learning via Importance Sampling
      • A Few Useful Things to Know About Machine Learning
  • Video Analytics
    • Index
      • Scaling Video Analytics on Constrained Edge Nodes
      • Focus: Querying Large Video Datasets with Low Latency and Low Cost
      • NoScope: Optimizing Neural Network Queriesover Video at Scale
      • Live Video Analytics at Scale with Approximation and Delay-Tolerance
      • Chameleon: Scalable Adaptation of Video Analytics
      • End-to-end Learning of Action Detection from Frame Glimpses in Videos
      • Short Summaries
  • Networking
    • Index
      • Salsify: Low-Latency Network Video through Tighter Integration between a Video Codec and a Transport
      • Learning in situ: a randomized experiment in video streaming
      • Short Summaries
  • Serverless
    • Index
      • Serverless Computing: One Step Forward, Two Steps Back
      • Encoding, Fast and Slow: Low-Latency Video Processing Using Thousands of Tiny Threads
      • SAND: Towards High-Performance Serverless Computing
      • Pocket: Elastic Ephemeral Storage for Serverless Analytics
      • Fault-tolerant and Transactional Stateful Serverless Workflows
  • Resource Disaggregation
    • Index
  • Edge Computing
    • Index
  • Security/Privacy
    • Index
      • Differential Privacy
      • Honeycrisp: Large-Scale Differentially Private Aggregation Without a Trusted Core
      • Short Summaries
  • Misc.
    • Index
      • Rate Limiting
      • Load Balancing
      • Consistency Models in Distributed System
      • Managing Complexity
      • System Design
      • Deep Dive into the Spark Scheduler
      • The Actor Model
      • Python Global Interpreter Lock
      • About Research and PhD
Powered by GitBook
On this page
  • What is Serverless Computing?
  • One Step Forward, Two Steps Back
  • Limitations of today's FaaS
  • Moving Forward:
  • Reference:

Was this helpful?

  1. Serverless
  2. Index

Serverless Computing: One Step Forward, Two Steps Back

http://cidrdb.org/cidr2019/papers/p119-hellerstein-cidr19.pdf

What is Serverless Computing?

At the core, serverless computing boils down to function-as-a-service(Faas).

  • Developers can register functions to run in the cloud, and declare when those functions should be triggered

  • Building applications on FaaS requires data management in both persistent and temporary storage, in addition to mechanisms to trigger and scale function execution.

  • The platform is not simply elastic, in the sense that humans or scripts can add and remove resources as needed, it is autoscaling: the workload automatically drives the allocation and deallocation of resources

  • The user is only billed for the computing resources used during function invocation

The difference between traditional cloud computing and serverless is that you, the customer who requires the computing, don’t pay for underutilized resources. Instead of spinning up a server in AWS for example, you’re just spinning up some code execution time. The serverless computing service takes your functions as input, performs logic, returns your output, and then shuts down. You are only billed for the resources used during the execution of those functions.

One Step Forward, Two Steps Back

By providing autoscaling, today’s FaaS offerings take a big step forward for cloud programming, offering a practically manageable, seemingly unlimited computing platform. Unfortunately, as we will see, today’s FaaS offerings also slide two major steps backward. First, they painfully ignore the importance of efficient data processing. Second, they stymie the development of distributed systems

In short, current FaaS solutions are attractive for simple workloads of independent tasks—be they embarrassingly parallel tasks embedded in Lambda functions or jobs to be run by the proprietary cloud services. However, use cases that involve stateful tasks have surprisingly high latency.

Limitations of today's FaaS

  • Limited Lifetimes: Typically, after 15 minutes, function invocations are shut down by the Lambda infrastructure

  • I/O Bottlenecks: Functions are critically reliant on network bandwidth to communicate, but current platforms from AWS, Google, and Azure provide limited bandwidth.

  • Communication Through Slow storage: Since functions are not directly network accessible, they must communicate via an intermediary service(e.g., S3).

  • No Specialized Hardware: FaaS offerings today only allow users to provision a time slice of a CPU hyperthread and some amount of RAM. Prediction serving relies on access to specialized hardware like GPUs, which are not available through AWS Lambda.

These constraints and some shortcomings in the standard library of Faas offerings, taken together, substantially limit the scope of feasible serverless applications.

  • Data-shipping architecture: we move data to the code, not the other way round. However, it is conventional wisdom that moving data is more expensive than moving code.

  • The high cost of function interactions stymies basic distributed computing, which is founded on protocols performing fine-grained communication between agents(e.g., leader election). With all communication transiting through storage, there is no real way for thousands (much fewer millions) of cores in the cloud to work together efficiently using current FaaS platforms other than via largely uncoordinated (embarrassing) parallelism.

  • Current FaaS offerings all run on a uniform virutal machine platform, which stymies hardware-accelerated software innovation.

Moving Forward:

The paper identifies some key challenges that remain in achieving a truly programmable environment for the cloud.

  • Fluid Code and Data Placement: logical separation of code and data with the ability of the infrastructure to physically co-locate when it makes sense, including function shipping.

  • Heterogeneous Hardware Support

  • Long-running, addressable virtual agents

  • Disorderly programming: The sequential metaphor of procedural programming will not scale to the cloud. Developers need languages that encourage code that works correctly in small, granular units— of both data and computation— that can be easily moved around across time and space.

  • A common intermediate representation (IR), that can serve as a compilation target from many languages

  • Service-level objectives & guarantees

  • A ‘cloud-programming aware’ security model.

Reference:

PreviousIndexNextEncoding, Fast and Slow: Low-Latency Video Processing Using Thousands of Tiny Threads

Last updated 4 years ago

Was this helpful?

LogoServerless computing: one step forward, two steps back | the morning paper