White Papers

Comparing Microsoft SQL Server and SC Series availability features
10 SQL Server High Availability and Disaster Recovery with Dell EMC SC Series | CML1119
Always On Availability Groups in synchronous mode operates similar to SC Series Synchronous Replication
in High Availability mode. Data is replicated synchronously. In the event the connection to the secondary
exceeds the timeout threshold, the primary continues to accept writes, without replicating to the secondary,
until the issue is resolved.
When using either synchronous replication mechanism there is an additional cost that is introduced due to the
nature of the feature. Since both require acknowledgment from the secondary location prior to being
committed at the primary, the round-trip latency from the primary to the secondary is added to the transaction.
This is probably the largest deterrent from using a synchronous data replication approach.
There may not be a 1:1 relationship between a business transaction and a database transaction. It completely
depends on application design. For example, examine a simple business transaction such as saving new
customer data. To implement this as a database transaction, it could be a single database insert with all the
information, or it could be a separate database insert for customer name, home address, street address,
home phone, work phone, cell phone, etc. Even in this simple example, the additional latency impact to the
application varies greatly depending on the implementation. A synchronous replication implementation,
regardless of the method chosen, should be tested to assess the actual cost.
2.2 Other considerations
Both Microsoft SQL Server and SC Series arrays offer several features that can be used to satisfy high
availability and disaster recovery requirements for SQL Server databases. Snapshots, mirroring, and
replication features for database protection are available in both products but with very different
implementations largely due to their layer in the architecture stack. This sometimes makes it difficult to do a
direct technical comparison of these features. In addition to comparing them on a feature level, there are
other factors that are worth considering.
2.2.1 Expertise
SQL Server HA and DR features can be complex and require expert level DBA skills to architect and
maintain. Not only is a dedicated DBA resource likely required, but the design and day to day troubleshooting
of HA and DR features are typically duties of a senior SQL Server DBA. This may increase the number of
staff and/or skillsets that are required to setup, maintain, and recover these environments. With SC Series,
configuring HA and DR features is a very simple process using the same tools and concepts that storage
administrators are already familiar with. While DBAs need to be aware of the process, this presents the
opportunity to offload this workload and allow them to focus on other tasks. If a question arises, CoPilot
support is available 24x7 to assist with questions or issues that may arise.
2.2.2 Licensing
Software licensing costs can be an important part of the architecture decision process and should be
considered. In large architectures with hundreds of database instances, licensing costs could be substantial. If
availability and/or recovery features are a determining factor in choosing a SQL Server version, it may be
more cost-effective to use array-based features instead. Examine the license costs for different architectures
to determine if there is a significant cost difference that may play a role in the architecture decision.
2.2.3 Compatibility
While Microsoft SQL Server contains a robust set of features for database high availability (HA) and disaster
recovery (DR), those features only apply to SQL Server databases. In a database application environment,
there are typically other components external to SQL Server such as file storage, configuration files, other