DewDB was built without TiKV, and that was a deliberate choice about who owns the hard parts of a distributed database. According to its author, Vivek, in a DEV Community post, DewDB’s own nodes handle storage, replication, leader election, failover, sharding, and data movement. There is no TiKV underneath and no separate coordinator. The alternative he rejected was a document and query layer sitting on top of another distributed database.
What “without TiKV” actually means
TiKV is a distributed key-value store that other systems commonly use as their storage layer. A database built on it typically delegates storage and much of the distributed behavior to that layer, while the database itself adds the query language, document model, and client-facing features. DewDB takes the opposite route. The process you deploy is the process that stores data, copies it to other machines, decides who leads, recovers from failures, and moves data between shards.
As an Amazon Associate I earn from qualifying purchases.
The author’s own framing of the rejected path makes the reasoning explicit: “Because then DewDB would become a document and query layer on top of another distributed database.” The design premise is stated just as directly: “I wanted the thing you deploy to also be the thing doing the replication, failover, and sharding.”
Free tools Windows power users keep installed
One-click scans. No signup required.
How DewDB is organized, as the article describes it
The article describes two levels of structure. Each level is simple to state, but each carries its own failure logic.
#1 Best Overall
Replicated groups with leaders and replicas
A DewDB replicated group consists of a leader and one or more replicas. If the leader dies, a replica can take over. Leader election and the decision about which replica is safe to promote are handled inside DewDB, not by an external coordinator.
Scale-out through shard groups
For horizontal scale, the article describes multiple shard groups. Each shard group has its own leader and replicas, so different shards can accept writes independently. Shard ownership and the movement of data between shard groups are also DewDB responsibilities, including the online shard migration the author lists among its features.
Rank #2
What DewDB has to implement itself
The author states that the design makes DewDB responsible for the following. Each item is a mechanism that a layered design would usually hand to its storage layer or coordinator:
- Leader elections
- Quorum logic, meaning deciding when enough replicas have agreed for a write or a promotion to count
- WAL (write-ahead log) recovery after a crash
- Replica repair, bringing a lagging or damaged replica back in line
- Shard ownership, tracking which group is responsible for which data
- Migration, moving shard data while the system keeps running
Owning these mechanisms means owning the bugs in them. A failover path that works in one scenario but not another is a database defect, not a dependency issue to hand off to someone else.
Rank #3
The trade-off in system ownership
The author’s stated preference is deployment simplicity: one integrated system that performs replication, failover, and sharding. The cost is that the team building DewDB carries the complexity that a separate distributed storage layer would otherwise absorb. The table below sets out the ownership question in general terms. It does not measure either approach.
| Concern | Integrated design (DewDB, per the article) | Layered design on a distributed store |
|---|---|---|
| Replication | Implemented by DewDB nodes | Provided by the underlying distributed store, which the database depends on |
| Failover and leader election | Implemented by DewDB nodes | Depends on how the underlying store and database divide the work; the article does not describe a specific division |
| Sharding and data movement | Implemented by DewDB, including online shard migration | Depends on the layers; not stated in the article |
| Separate coordinator or storage cluster to deploy | None, according to the author | A separate distributed storage layer must be deployed and operated |
| Team that implements and operates quorum logic, recovery, and repair | The DewDB project | Split between the database project and the storage layer’s maintainers |
| Measured performance or reliability comparison | Not stated in the article | Not stated in the article |
When this ownership model fits
The integrated approach is most sensible when the following conditions hold. If several do not, a layered design may be the better fit:
Rank #4
- You want a single artifact to deploy, configure, and troubleshoot rather than a database plus a distributed storage cluster.
- Your team is willing to own election, quorum, recovery, and migration code, and to test failure paths thoroughly.
- You value control over how failover and shard movement behave, instead of depending on another project’s release schedule.
- You can accept that the database is less battle-tested than a widely deployed storage layer, unless you have evidence to the contrary for your workload.
What the source does and does not establish
The design rationale comes from the author’s own article, “Why I Built DewDB Without TiKV,” on DEV Community: https://dev.to/viveke22/why-i-built-dewdb-without-tikv-16bc. The accessible excerpt shows “Posted on Sep 26” but does not establish a publication year, so no year is attached here. Vivek’s role beyond authorship is not described.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallThe article is a statement of intent and architecture. It does not establish:
- Performance figures. No benchmark or throughput or latency measurement is given.
- Production readiness, or the maturity of the failover and migration code.
- Supported platforms. The article includes Windows installation commands and a GitHub link, but those commands are not reproduced or verified here.
- Comparative reliability against TiKV-based systems or any other database.
- Whether the feature list is current. The article’s feature snapshot covers documents, queries, secondary indexes, replication, failover, sharding, change streams, and online shard migration. Treat it as the author’s snapshot at the time of writing, not an independently audited list.
The fair reading is that the article explains why DewDB was built as an integrated system and what that choice costs. Whether the implementation handles those costs well is a separate question that the article does not answer.
No commercial relationship, pricing, or availability program for DewDB is described in the source.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
