Cloud Cover AMA | Secure Multi-Cloud:
Join us for Ecclesia #25, our next Züs Cloud Cover AMA, on October 12th, at 9 AM PST. With the Active Set in the midst of redeployment, Saswata will discuss the next steps and answer your questions about what lies ahead. Share your inquiries and insights beforehand on our Discord or Telegram. It is a pivotal time for our community, and your participation is highly valued. See you there! Read on to learn more about how Züs is the Secure Multi-Cloud.
As well, we are grateful for the support on Product Hunt! Your enthusiasm has sparked interest from other projects and the Product Hunt team. If you have not already, please click “Notify me.” It is a simple step, and you can use your Google account to sign in. Your involvement makes a big difference to us!
Storm of the Week:
Secure Multi-Cloud with Züs
Züs is revolutionizing decentralized storage. Imagine your data, not trapped in a vulnerable server, but fragmented, encrypted, and spread across the globe, reconstructing only under your command. Züs delivers more than just storage; it is a passport to data freedom and security, granting you the power to choose only the best providers. It is secure, fast, and boundless.
As we near the Mainnet’s launch, we are not just safeguarding bytes; we are sending them into a world full of new possibilities.
Stay connected with Züs — where the next chapter of storage technology is being written, byte by byte.
Last week, the blockchain team continued investigating the MPT (Merkelized Patricia Tree) state growth issue, leading to the implementation of several new fixes aimed at optimizing the MPT state size. Numerous backend issues were resolved across various repositories.
For further details, see below:
In continuation of the challenges addressed in our previous report, the team conducted in-depth research on how RocksDB manages the SST (Sorted String Table) log files and implemented metrics to monitor the total key numbers within RocksDB. It was discovered that RocksDB does not immediately reclaim disk space following the deletion of items. Essentially, all actions such as insert, delete, update are logged as new entries, which do not get removed from the disk until certain conditions are fulfilled. Based on these findings, the team adjusted the RocksDB configuration to reduce the time taken to delete obsolete files from 6 hours to 10 minutes. This modification successfully kept disk usage at a reasonable size. Furthermore, the state prune period of 0chain was reduced from 30 seconds to 7 seconds to minimize the retention of unnecessary deleted keys in MPT. Detailed changes can be viewed in PR 2837.
The team also rolled out several updates.
Beyond these configuration changes, the team also rolled out several updates in smart contracts to liberate resources. This involved removing the allocation from MPT when it’s either finalized or canceled, in addition to removing blobbers/validators from MPT upon their termination or shutdown. In the past, finalized/canceled allocations remained in MPT, occupying space unnecessarily. The same held true for blobbers and validators. With the latest fixes in PRs 2829 and 2841, all such redundancies are now effectively cleared from the MPT, freeing up significant space.
Here are the core backend PRs that have been merged:
- PR 2820: A snapshot update challenge was fixed.
- PR 2811: A node recovery conductor testing workflow was added.
- PR 2812: The maximum file size on blobber was limited.
- PR 2826: An issue with the challenge pool was resolved.
- PR 2814: LFB is now loaded from the state DB on start for sharders, storing LFB round info, including round number and block hash to state DB. If it exists, it’s used as LFB on starting; otherwise, the process iterates from the last one from round stores as before.
- PR 2836: A benchmark test of storage.challenge_response was fixed.
- PR 2837: Fixes were applied regarding MPT state size growth.
- PR 2838: A challenge generated period config was added.
- PR 2842: An issue with third party extendability was fixed.
- PR 2847: Additional LFB loading on sharders was corrected.
- PR 2849: Unused input was removed.
- PR 2829: Allocation was removed from MPT when it’s finalized/canceled. Write pool tokens were refunded to the allocation owner automatically when it’s finalized/canceled.
- PR 2841: Blobber/validator was deleted from MPT when it’s killed/shutdown.
- PR 1245: Concurrency control was introduced in gosdk.
- PR 1247: Stream tests were fixed.
- PR 1246: Share consensus was corrected.
- PR 1240: List optimization was performed.
- PR 1231: A sharder keep list was added.
- PR 1277: A challenge timing issue was resolved.
- PR 1258: L2 caching for allocation was added.
- PR 1284: Optimization of GetRefs was carried out, along with corrections and the addition of indexes on the ref table.
- PR 1285: Finalization issues were fixed.
A multi-cloud solution that is more secure
Züs provides a powerful, secure and reliable platform for organizations who are looking to make the switch to decentralized storage. It leverages the strengths of multiple cloud providers alongside blockchain technology to ensure better security and reliability at scale. Thank you for reading about how Züs helps secure multi-cloud operations. Join the Züs AMA tomorrow – where we will answer your questions.