Bluzelle, a
brainchild of Pavel Bains and Neeraj Murarka was formed in July 2014. Bluzelle created
a technology stack that can abstract the complexities of managing blockchain or
decentralized apps. Bluzelle aims to be the number one data layer for Web 3.0
which provides a decentralized data network that is secure, tamper-proof,
scalable, and available and that has greater privacy barring all forms of disadvantages
experienced by many blockchains. Bluzelle has been in existence since 2014 and
was awarded a prize of Technology Pioneer by the World Economic Forum in 2017. We
need to look into the technology uniqueness of bluzelle to understand why it
was awarded such an honoring prize.
On bluzelle database
like in other databases, the four basic functions of database acronym CRUD that
is Create, Read, Update, and Delete are mapped to their related methods. To use
the bluzelle database, a password that must be protected by the user is
required. Note that this password is stored locally on the user’s device and
does not get transmitted through bluzelle’s network.
All key-value pair data on
the network are encrypted using the user’s password as the initialization
vector in AES symmetric key encryption. This encryption of data on bluzelle
ensures maximum security for data stored on its database.
More so ,In order to ensure
scalability and prevent over clogging of the database as data storage expands
and leads to slow and expensive retrieval of data(A major issue on ethereum on
which ethereum2 hinges),bluzelle adopts the strategy of sharding. SHARD stands
for System for Highly Available Replicated Data. In simple terms, sharding is
the partitioning of a data storage logical paths.The smallest unit of a
bluzelle database is a logical shard that contains units of data from various
users. However this shard shares a unique but uniform partition key as an
identifier which allows the shard to access and retrieve its information in the
future.
In order
to be able to map the key in a key-value pair in a NoSQL table to the Id of swarm
(A swarm on bluzelle is a specific collection of nodes on the bluzelle
network). Once the Id is found, bluzelle makes use of the Kademila hashing to
reach that swarm even if it is not running at the moment.
Bluzelle
makes use of Kademila hashing to simplify the process of the recognition of
every other swarm by a specific node. Kademila is a highly efficient peer to
peer distributed hash table designed to make use of a specific symmetric and
geometric properties of the bitwise XOR function. Kademila’s table in the
network needs only to know how to get to other leaf swarms. This means no
matter how large the network becomes, every node will be able to reach other
related swarms which is a great solution to the problem of scalability.
Decentralized storage
systems like ethereum are bedeviled with redundancy and efficiency because in
order to ensure security, the ethereum network had to trade off efficiency (data
storage duplication) in that every ethereum node must keep a copy of a set of
data and such data must be confirmed by the entire network before retrieval
leading to high cost of data storage and slow data retrieval process. The
system then suffers from vertical inscalability. To prevent this problem,
bluzelle adopts the partial replication strategy.
The
strategy of partial replication ensures that not all nodes in the network has a
copy of a particular data set, rather only the nodes within the swarm delegated
to such data replicate it. Bluzelle stores data only on specific swarms which
are subsets of the entire network’s nodes. This ensures both vertical and
horizontal scalability, reduce cost of data storage, increase data storage
efficiency as well as ensure security. Partial replication is the same strategy
adopted by Content Delivery Network (CDN).
Bluzelle still ensures security because a
swarm is a small collection of nodes on which data are stored. So, by having
the same data stored on different hardwares in various geographical locations,
it means that such data are secured and still the principle of load balancing
makes the network fast.
Continue to part 2 here
Comments
Post a Comment