Skip to main content
Version: 0.50

ADR-065: Store V2

Changelog​

  • Feb 14, 2023: Initial Draft (@alexanderbez)

Status​

DRAFT

Abstract​

The storage and state primitives that Cosmos SDK based applications have used have by and large not changed since the launch of the inaugural Cosmos Hub. The demands and needs of Cosmos SDK based applications, from both developer and client UX perspectives, have evolved and outgrown the ecosystem since these primitives were first introduced.

Over time as these applications have gained significant adoption, many critical shortcomings and flaws have been exposed in the state and storage primitives of the Cosmos SDK.

In order to keep up with the evolving demands and needs of both clients and developers, a major overhaul to these primitives are necessary.

Context​

The Cosmos SDK provides application developers with various storage primitives for dealing with application state. Specifically, each module contains its own merkle commitment data structure -- an IAVL tree. In this data structure, a module can store and retrieve key-value pairs along with Merkle commitments, i.e. proofs, to those key-value pairs indicating that they do or do not exist in the global application state. This data structure is the base layer KVStore.

In addition, the SDK provides abstractions on top of this Merkle data structure. Namely, a root multi-store (RMS) is a collection of each module's KVStore. Through the RMS, the application can serve queries and provide proofs to clients in addition to provide a module access to its own unique KVStore though the use of StoreKey, which is an OCAP primitive.

There are further layers of abstraction that sit between the RMS and the underlying IAVL KVStore. A GasKVStore is responsible for tracking gas IO consumption for state machine reads and writes. A CacheKVStore is responsible for providing a way to cache reads and buffer writes to make state transitions atomic, e.g. transaction execution or governance proposal execution.

There are a few critical drawbacks to these layers of abstraction and the overall design of storage in the Cosmos SDK:

  • Since each module has its own IAVL KVStore, commitments are not atomic
    • Note, we can still allow modules to have their own IAVL KVStore, but the IAVL library will need to support the ability to pass a DB instance as an argument to various IAVL APIs.
  • Since IAVL is responsible for both state storage and commitment, running an archive node becomes increasingly expensive as disk space grows exponentially.
  • As the size of a network increases, various performance bottlenecks start to emerge in many areas such as query performance, network upgrades, state migrations, and general application performance.
  • Developer UX is poor as it does not allow application developers to experiment with different types of approaches to storage and commitments, along with the complications of many layers of abstractions referenced above.

See the Storage Discussion for more information.

Alternatives​

There was a previous attempt to refactor the storage layer described in ADR-040. However, this approach mainly stems on the short comings of IAVL and various performance issues around it. While there was a (partial) implementation of ADR-040, it was never adopted for a variety of reasons, such as the reliance on using an SMT, which was more in a research phase, and some design choices that couldn't be fully agreed upon, such as the snap-shotting mechanism that would result in massive state bloat.

Decision​

We propose to build upon some of the great ideas introduced in ADR-040, while being a bit more flexible with the underlying implementations and overall less intrusive. Specifically, we propose to:

  • Separate the concerns of state commitment (SC), needed for consensus, and state storage (SS), needed for state machine and clients.
  • Reduce layers of abstractions necessary between the RMS and underlying stores.
  • Provide atomic module store commitments by providing a batch database object to core IAVL APIs.
  • Reduce complexities in the CacheKVStore implementation while also improving performance[3].

Furthermore, we will keep the IAVL is the backing commitment store for the time being. While we might not fully settle on the use of IAVL in the long term, we do not have strong empirical evidence to suggest a better alternative. Given that the SDK provides interfaces for stores, it should be sufficient to change the backing commitment store in the future should evidence arise to warrant a better alternative. However there is promising work being done to IAVL that should result in significant performance improvement [1,2].

Separating SS and SC​

By separating SS and SC, it will allow for us to optimize against primary use cases and access patterns to state. Specifically, The SS layer will be responsible for direct access to data in the form of (key, value) pairs, whereas the SC layer (IAVL) will be responsible for committing to data and providing Merkle proofs.

Note, the underlying physical storage database will be the same between both the SS and SC layers. So to avoid collisions between (key, value) pairs, both layers will be namespaced.

State Commitment (SC)​

Given that the existing solution today acts as both SS and SC, we can simply repurpose it to act solely as the SC layer without any significant changes to access patterns or behavior. In other words, the entire collection of existing IAVL-backed module KVStores will act as the SC layer.

However, in order for the SC layer to remain lightweight and not duplicate a majority of the data held in the SS layer, we encourage node operators to keep tight pruning strategies.

State Storage (SS)​

In the RMS, we will expose a single KVStore backed by the same physical database that backs the SC layer. This KVStore will be explicitly namespaced to avoid collisions and will act as the primary storage for (key, value) pairs.

While we most likely will continue the use of cosmos-db, or some local interface, to allow for flexibility and iteration over preferred physical storage backends as research and benchmarking continues. However, we propose to hardcode the use of RocksDB as the primary physical storage backend.

Since the SS layer will be implemented as a KVStore, it will support the following functionality:

  • Range queries
  • CRUD operations
  • Historical queries and versioning
  • Pruning

The RMS will keep track of all buffered writes using a dedicated and internal MemoryListener for each StoreKey. For each block height, upon Commit, the SS layer will write all buffered (key, value) pairs under a RocksDB user-defined timestamp column family using the block height as the timestamp, which is an unsigned integer. This will allow a client to fetch (key, value) pairs at historical and current heights along with making iteration and range queries relatively performant as the timestamp is the key suffix.

Note, we choose not to use a more general approach of allowing any embedded key/value database, such as LevelDB or PebbleDB, using height key-prefixed keys to effectively version state because most of these databases use variable length keys which would effectively make actions likes iteration and range queries less performant.

Since operators might want pruning strategies to differ in SS compared to SC, e.g. having a very tight pruning strategy in SC while having a looser pruning strategy for SS, we propose to introduce an additional pruning configuration, with parameters that are identical to what exists in the SDK today, and allow operators to control the pruning strategy of the SS layer independently of the SC layer.

Note, the SC pruning strategy must be congruent with the operator's state sync configuration. This is so as to allow state sync snapshots to execute successfully, otherwise, a snapshot could be triggered on a height that is not available in SC.

State Sync​

The state sync process should be largely unaffected by the separation of the SC and SS layers. However, if a node syncs via state sync, the SS layer of the node will not have the state synced height available, since the IAVL import process is not setup in way to easily allow direct key/value insertion. A modification of the IAVL import process would be necessary to facilitate having the state sync height available.

Note, this is not problematic for the state machine itself because when a query is made, the RMS will automatically direct the query correctly (see Queries).

Queries​

To consolidate the query routing between both the SC and SS layers, we propose to have a notion of a "query router" that is constructed in the RMS. This query router will be supplied to each KVStore implementation. The query router will route queries to either the SC layer or the SS layer based on a few parameters. If prove: true, then the query must be routed to the SC layer. Otherwise, if the query height is available in the SS layer, the query will be served from the SS layer. Otherwise, we fall back on the SC layer.

If no height is provided, the SS layer will assume the latest height. The SS layer will store a reverse index to lookup LatestVersion -> timestamp(version) which is set on Commit.

Proofs​

Since the SS layer is naturally a storage layer only, without any commitments to (key, value) pairs, it cannot provide Merkle proofs to clients during queries.

Since the pruning strategy against the SC layer is configured by the operator, we can therefore have the RMS route the query SC layer if the version exists and prove: true. Otherwise, the query will fall back to the SS layer without a proof.

We could explore the idea of using state snapshots to rebuild an in-memory IAVL tree in real time against a version closest to the one provided in the query. However, it is not clear what the performance implications will be of this approach.

Atomic Commitment​

We propose to modify the existing IAVL APIs to accept a batch DB object instead of relying on an internal batch object in nodeDB. Since each underlying IAVL KVStore shares the same DB in the SC layer, this will allow commits to be atomic.

Specifically, we propose to:

  • Remove the dbm.Batch field from nodeDB
  • Update the SaveVersion method of the MutableTree IAVL type to accept a batch object
  • Update the Commit method of the CommitKVStore interface to accept a batch object
  • Create a batch object in the RMS during Commit and pass this object to each KVStore
  • Write the database batch after all stores have committed successfully

Note, this will require IAVL to be updated to not rely or assume on any batch being present during SaveVersion.

Consequences​

As a result of a new store V2 package, we should expect to see improved performance for queries and transactions due to the separation of concerns. We should also expect to see improved developer UX around experimentation of commitment schemes and storage backends for further performance, in addition to a reduced amount of abstraction around KVStores making operations such as caching and state branching more intuitive.

However, due to the proposed design, there are drawbacks around providing state proofs for historical queries.

Backwards Compatibility​

This ADR proposes changes to the storage implementation in the Cosmos SDK through an entirely new package. Interfaces may be borrowed and extended from existing types that exist in store, but no existing implementations or interfaces will be broken or modified.

Positive​

  • Improved performance of independent SS and SC layers
  • Reduced layers of abstraction making storage primitives easier to understand
  • Atomic commitments for SC
  • Redesign of storage types and interfaces will allow for greater experimentation such as different physical storage backends and different commitment schemes for different application modules

Negative​

  • Providing proofs for historical state is challenging

Neutral​

  • Keeping IAVL as the primary commitment data structure, although drastic performance improvements are being made

Further Discussions​

Module Storage Control​

Many modules store secondary indexes that are typically solely used to support client queries, but are actually not needed for the state machine's state transitions. What this means is that these indexes technically have no reason to exist in the SC layer at all, as they take up unnecessary space. It is worth exploring what an API would look like to allow modules to indicate what (key, value) pairs they want to be persisted in the SC layer, implicitly indicating the SS layer as well, as opposed to just persisting the (key, value) pair only in the SS layer.

Historical State Proofs​

It is not clear what the importance or demand is within the community of providing commitment proofs for historical state. While solutions can be devised such as rebuilding trees on the fly based on state snapshots, it is not clear what the performance implications are for such solutions.

Physical DB Backends​

This ADR proposes usage of RocksDB to utilize user-defined timestamps as a versioning mechanism. However, other physical DB backends are available that may offer alternative ways to implement versioning while also providing performance improvements over RocksDB. E.g. PebbleDB supports MVCC timestamps as well, but we'll need to explore how PebbleDB handles compaction and state growth over time.

References​