Versions Compared

Old Version 244

changes.mady.by.user Stephen M. Rumble

Saved on

compared with

New Version Current

changes.mady.by.user John Ousterhout

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

What is RAMCloud?

The RAMCloud project is creating a new class of super-high-speed storage , based entirely in DRAM, that is 2-3 orders of magnitude faster than existing storage systems. If successful, it will enable new applications that manipulate large-scale datasets much more intensively than has ever been possible before. In addition, we think for large-scale datacenter applications. It is designed for applications in which a large number of servers in a datacenter need low-latency access to a large durable datastore. RAMCloud offers the following properties:

  • Low latency: RAMCloud keeps all data in DRAM at all times, so applications can read RAMCloud objects remotely over a datacenter network in as little as 5μs. Writes take less than 15μs. Unlike systems such as memcached, applications never have to deal with cache misses or wait for disk/flash accesses. As a result, RAMCloud storage is 10-1000x faster than other available alternatives.
  • Large scale: RAMCloud aggregates the DRAM of thousands of servers to support total capacities of 1PB or more.
  • Durability: RAMCloud replicates all data on nonvolatile secondary storage such as disk or flash, so no data is lost if servers crash or the power fails. One of RAMCloud's unique features is that it recovers very quickly from server crashes (only 1-2 seconds) so the availability gaps after crashes are almost unnoticeable. As a result, RAMCloud combines the durability of replicated disk with the speed of DRAM. If you have used memcached, you have probably experienced the challenges of managing a second durable storage system and maintaining consistency between it and memcached. With RAMCloud, there is no need for a second storage system.
  • Powerful data model: RAMCloud's basic data model is a key-value store, but we have extended it with several additional features, such as:
    • Multiple tables, each with its own key space.
    • Transactional updates that span multiple objects in different tables.
    • Secondary indices.
    • Strong consistency: unlike other NoSQL storage systems, all updates in RAMCloud are consistent, immediately visible, and durable.
  • Easy deployment: RAMCloud is a software package that runs on commodity Intel servers with the Linux operating system. RAMCloud is available freely in open source form.

From a practical standpoint, RAMCloud enables a new class of applications that manipulate large data sets very intensively. Using RAMCloud, an application can combine tens of thousands of items of data in real time to provide instantaneous responses to user requests.  Unlike traditional databases, RAMCloud scales to support very large applications, while still providing a high level of consistency. We believe that RAMCloud, or something like it, will become the primary storage system for structured data in cloud computing environments such as Amazon's AWS and or Microsoft's Azure.

The role of DRAM in storage systems has been increasing rapidly in recent years, driven by the needs of large-scale Web applications. These applications manipulate very large datasets with an intensity that cannot be satisfied by disks alone. As a result, applications are keeping more and more of their data in DRAM. For example, large-scale caching systems such as memcached are being widely used (in 2009 Facebook used a total of 150 TB of DRAM in memcached and other caches for a database containing 200 TB of disk storage), and the major Web search engines now keep their search indexes entirely in DRAM.

Although DRAM's role is increasing, it still tends to be used in limited or specialized ways. In most cases DRAM is just a cache for some other storage system such as a database; in other cases (such as search indexes) DRAM is managed in an application-specific fashion. It is difficult for developers to use DRAM effectively in their applications; for example, the application must manage consistency between caches and the backing storage. In addition, cache misses and backing store overheads make it difficult to capture DRAM's full performance potential.

Our goal for RAMCloud is to create a general-purpose storage system that makes it easy for developers to harness the full performance potential of large-scale DRAM storage. It keeps all data in DRAM all the time, so there are no cache misses. RAMCloud storage is durable and available, so developers need not manage a separate backing store. RAMCloud is designed to scale to thousands of servers and hundreds of terabytes of data while providing uniform low-latency access to all machines within a large datacenter.

As of Fall 2011, we had initial implementations of many of the components of RAMCloud and the system runs well enough to use it for simple tests. On our 60-node test cluster we are able to perform remote reads of 100-byte objects in about 5 microseconds, and an individual server can process more than 800,000 small read requests per second. The basic crash recovery mechanism is running, and RAMCloud can recover 35 GB of memory from a failed server in about 1.6 seconds.

The RAMCloud project is still young, so there are many interesting research issues still to explore, such as the following:

...

We have built the system not as a research prototype, but as a production-quality software system, suitable for use by real applications.

RAMCloud is also interesting from a research standpoint. Its two most important attributes are latency and scale. The first goal is to provide the lowest possible end-to-end latency for applications accessing the system from within the same datacenter. We currently achieve latencies of around 5μs for reads and 15μs for writes, but hope to improve these in the future. In addition, the system must scale, since no single machine can store enough DRAM to meet the needs of large-scale applications. We have designed RAMCloud to support at least 10,000 storage servers; the system must automatically manage all the information across the servers, so that clients do not need to deal with any distributed systems issues. The combination of latency and scale has created a large number of interesting research issues, such as how to ensure data durability without sacrificing the latency of reads and writes, how to take advantage of the scale of the system to recover very quickly after crashes, how to manage storage in DRAM, and how to provide higher-level features such as secondary indexes and

...

multiple-object transactions

...

without sacrificing the latency or scalability

...

of the system.

...

Our solutions to these problems are described in a series of technical papers.

The RAMCloud project was based in the Department of Computer Science at Stanford University. The project is no longer active and the students working on RAMCloud have graduated, so we cannot provide support for anyone wishing to use RAMCloud.

Learning About RAMCloud

General information about RAMCloud, such as talks and papers. Much of the information here is related to the research aspects of the project, as opposed to information on how to use RAMCloud.

How to Deploy and Use RAMCloud

RAMCloud has now reached a level of maturity where it is suitable for production use with real applications.  The links below provide information on how to set up a RAMCloud cluster and on the RAMCloud APIs for applications.

RAMCloud Performance

Measurements of RAMCloud performance, as well as comparisons between RAMCloud and other systems.

...

Resources

...

Information for RAMCloud Developers

Information for people who are working on the RAMCloud code base; it is intended primarily for the internal use of the RAMCloud team at Stanford, but may be useful to other people as well.

...

The RAMCloud Test Cluster

...

Information

...

RAMCloud Cluster

...

about the cluster we use for RAMCloud testing at Stanford. Unfortunately not all of this information is completely up to date.

Informational

Current work

Old Topics

New Cluster

  • ATOM Cluster :  Micro Modular Server  Cluster – 132 ATOM servers

Design Notes

These documents were used at various points in the project to record our early ideas about various parts of the system. Most of these pages are now out of date (they typically are not updated once serious coding begins) but they may still provide useful background information as well as alternatives that we considered. Entries below are in reverse chronological order (most recent design notes first).

Project History, Schedules, Milestones

Miscellaneous Topics

Ideas for Future Work

Related Topics

Miscellaneous Topics

Personal Wikis