SQL Server Dba Responsibilities & Implications of Integration
SQL Server Dba Responsibilities & Implications of Integration
▲
SQL SERVER DBA
RESPONSIBILITIES
& IMPLICATIONS
OF INTEGRATION
SearchSQLServer
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
Home
“D
In-Memory SQL
Server: The Implica-
tions of Integration
PA G E 2 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
Defining Half a Dozen SQL Server database administration can be a complex and stressful job.
Daily SQL Server DBA
Responsibilities Database administrators’ responsibilities cover the performance, integrity
and security of business data and SQL Server databases. To fulfill their duties
In-Memory SQL
Server: The Implica- and to make business data available to its users, database administrators have
tions of Integration
to perform routine DBA checks on their SQL Servers to monitor their status.
So, what critical aspects of SQL Server should all DBAs include in their
daily checklist? Here are six daily DBA responsibilities that every SQL Server
manager should perform.
PA G E 3 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
PA G E 4 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
PA G E 5 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
Home
In-Memory SQL
Server: The Implica-
tions of Integration
PA G E 6 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
PA G E 7 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
Home
In-Memory SQL
Server: The Implica-
tions of Integration
PA G E 8 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
PA G E 9 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
PA G E 1 0 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
Defining Half a Dozen It’s easy to be misled by terminology. When I first heard the term “in-memory
Daily SQL Server DBA
Responsibilities database” applied to SQL Server, I made the same mistake many other people
have. I thought, “How’s that different from an instance of SQL Server with
In-Memory SQL
Server: The Implica- tons of RAM and the buffer pooling for the program turned up to 11?” But in-
tions of Integration
memory SQL Server -- a major change, codenamed Hekaton and planned for
the next iteration of the product -- isn’t like that. And why it isn’t like that has
implications for how such an instance of SQL Server would fit in with the rest
of your setup.
First, the basics. The label conveys the fundamental idea behind an in-
memory database pretty accurately: It’s a database system where both the
engine and as much of the data as possible are stored directly in RAM. Of
course, this dramatically increases the speed of transactions, but this is only
possible in a couple circumstances:
PA G E 1 1 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
1. The database and engine are small enough to fit into RAM by default
2. The system contains enough RAM to hold both database and engine
With memory getting cheaper all the time and the average server sporting
Home
more of it, scenario #2 has been happening a lot more often than scenario #1
Defining Half a Dozen ever did. The sheer size of almost any professional SQL Server deployment all
Daily SQL Server DBA
Responsibilities but guarantees the former rarely happens.
But it isn’t just memory that’s grown cheaper, faster and more plentiful.
In-Memory SQL
Server: The Implica- CPUs have also become dramatically faster and more parallel, and in order
tions of Integration
to keep that silicon from just sitting around doing nothing, more of what the
database does is being moved into memory whenever possible.
Because of these issues, the forthcoming in-memory enhancements to SQL
Server are being engineered to take advantage of all these trends. In a TechNet
blog post entitled “The coming in-memory database tipping point,” David
Campbell lays out some of these changes, including how the database itself is
stored differently in memory (via a columnar, rather than a row-based, model).
Microsoft has in fact already adopted some of these methods for the
PowerPivot add-on for Microsoft Excel. “In SQL Server 2012,” Campbell
explains, “this ships as the xVelocity in-memory analytics engine as part of SQL
PA G E 1 2 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
Server Analysis Services.” In the long run, these components are going to be
used in other parts of SQL Server, so that the power unleashed by in-memory
databases can be used outside of the highly vertical solutions that currently
exploit it most, such as data warehousing or analysis.
Home
This brings up the first question I’ve been building towards: Would all of
Defining Half a Dozen your SQL Server installations need to be upgraded to take full advantage of
Daily SQL Server DBA
Responsibilities in-memory processing? The short answer is probably, but with three caveats:
In-Memory SQL
Server: The Implica- 1. IN-MEMORY DATABASE SOLUTIONS DO REQUIRE SOME RE-ARCHITECTING OF
tions of Integration
YOUR EXISTING DATABASE
The columnar, rather than row-based, layout in-memory databases use will
require some work on your part to set up. In one of the case studiesMicrosoft
released to talk about SQL Server 2012’s existing in-memory features, they
noted that while some changes to the database were needed, they required little
more than “chang[ing] some metadata values.” Depending on the way your
data is already set up, the amount of work may be minimal, but don’t assume
it’s going to be zero.
PA G E 1 3 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
2. THE LARGER THE DATABASE, AND THE MORE IT FITS THE SCENARIO, THE
BIGGER THE PAYOFF
You’re likely to see the biggest improvements -- the ones most worth the
effort of retooling everything to use in-memory processing -- from the largest
Home
workloads that are analysis- and data-warehouse-centric. Campbell notes
Defining Half a Dozen that the columnar structure of in-memory databases “are not optimal for
Daily SQL Server DBA
Responsibilities transaction processing workloads,” the conventional CRUD (create, read,
update, delete) workloads that most of us learned about in DBMS 102. He does
In-Memory SQL
Server: The Implica- note that in time, in-memory technology will be expanded to include these more
tions of Integration
conventional scenarios, but it won’t happen right away.
PA G E 1 4 O F 1 5 SPONSORED BY
SQL SERVER DBA RESPONSIBILITIES & IMPLICATIONS OF INTEGRATION
PA G E 1 5 O F 1 5 SPONSORED BY