How Cosmos DB Handles Data Consistency - PowerPoint PPT Presentation

About This Presentation
Title:

How Cosmos DB Handles Data Consistency

Description:

Cosmos DB users with limited percentage will use this approach for data consistency in the real world. – PowerPoint PPT presentation

Number of Views:50

less

Transcript and Presenter's Notes

Title: How Cosmos DB Handles Data Consistency


1
How Cosmos DB Handles Data Consistency
Cosmos DB users with limited percentage will use
this approach for data consistency in the real
world. Advantages of three alternative
consistency models will be used instead. Turing
Award Winner Leslie Lamport is based on the work.
Thereby creating a database for managing the
real life situations were the foundations and
they deliver shared applications without the
traditional consistency model penalties.
Staleness is bounded by the first alternative
consistency model that offer a point that there
is a sync between reads and writes. There is no
guarantee before it but after it the latest
version is always accessed by you. Either a
number of versions or time interval you will
define the boundary. Everything is consistent
outside the boundary and within it there is no
surety of a read returning latest data. An
element of strong consistency is a store while
offering you low latency and the choice of
global sharing and higher reliability. If you
want to be sure of all the consistent reads then
you can use this model. The writes are also
considered to be fast. If you read it in the
region that is written then you can get the data
that is correct. Session consistency, second
alternative consistency model, works well when
there is a read and write drives from a client
app. The own writes of a client can be read by
them, and across the rest of the network the data
replication takes place. You have low latency of
data access with this way and you know that you
will fail over over a period of time and in any
Azure region your application will run. A third
alternative consistency model with prefix of
consistency in Cosmos DB has been added by
Microsoft. The eventual consistency speed has
added predictability with Consistency prefix.
The latest write might not be seen by you when
you read the data but your reads will never be
out of order. It is both fast and predictable
useful feature. Your client will be able to see
after Write A, then B, and then C, only A or A
and B but never A and C. The Cosmos DB regions
will mix on A, B, and C, offering you reliability
and speed. From the competition it is a very
different beast when Cosmos DB is concerned.
There are some NoSQl offers from the limited form
of shared access, but they target at redundancy
offers and recovery from disasters. There are
similar features offered by the Google
spanner but is possible only
2
datacenters in a single region. If your target
audience is only US or EU then you might be fine
with the work but for a global reach there are
more cloud services. Strong consistency in low
latency is a good option but has less value when
data replication that is cross-regional becomes
a major bottleneck. CosmosDB relies on
applications and that is what your type of
consistency choice is all about. Is your target
about reading or writing the data? Usage of data?
There are advantages and disadvantages of each
consistency model and you need to take into
account before choice preferences carefully. A
good place to start, session consistency for most
app-centric data. It is good to test with
various choices when you dont require immediate
access of data globally. For more information
join the DBA Course to make your career in this
field. Stay connected to CRB Tech for more
technical optimization and other updates and
information.
Write a Comment
User Comments (0)
About PowerShow.com