Harry Tolsma

SQL Server news from connect start to resonate

Hi all,

A month ago at the Connect event in New York Microsoft did some anouncements on SQL Server vNext and Azure Data Services.
It’s good to see the traction it’s getting!

I work with a lot of partners on implementing Microsoft stuff and the partners are very keen in getting the latest info on new capabilities.
Also they enlist in new online and in classroom courses on the new stuff. So let’s take a quick look:

  • Azure Data Lake store and analytics went live in Novmeber, This is a no limits data lake with on-demand analytics that instantly scales to your needs. With petabyte size files and trillions of objects and parallel processing!
  • R Server for Azure HD Insight for Largest R compatible parallel analytics ML library with terabyte-scale machine learning!
  • SQL Server on Linux public preview is released. Check in out on a linux box!
  • Moving selected Enterprise Edition capabilities to SQL Server Standard like in memory and Polybase.

And last but not least: connectivity to all top programming languages!

 

SQL Server op Linux? Ja dus!

Als iemand een paar jaar terug tegen mij had gezegd dat Microsoft Multi platform zou gaan en dat ook SQL Server op Linux zou uitkomen, dan had ik hem voor gek verklaard. Echter sinds de komst van Satya Nadella als CEO en mensen als Scott Guthrie (CVP voor Cloud & Enterprise) waait er een frisse wind door Microsoft!

En als Microsoft employee voelde ik natuurlijk al wat aankomen, maar het is toch altijd weer top om het nieuws dan nu in de buitenwereld te zien landen: Gister heeft Microsoft de Linux versie van SQL Server geannonceerd in preview. Dit betreft voorlopig alleen het RDBMS en niet de overige SQL tools als SSRS en SSAS, maar een mooie start om nu ook op Linux de (volgens Gartner) de leading database te kunnen runnen!

Ik zou het leuk vinden als jullie na het testen van deze versie op het blog zouden willen reageren met de eerste indruk van de performance van deze preview versie.

De blogpost van Scott over de annoncering: http://blogs.microsoft.com/blog/2016/03/07/announcing-sql-server-on-linux/

SQL groeten,

Harry

SQL Server 2005 is End-of-life in april volgend jaar

Voor iedere DBA-er en database manager is het belangrijk om alvast voor te sorteren op het feit dat het support van Microsoft op SQL 2005 eindigt in april volgend jaar.

Dit lijkt ver weg, maar omdat je als beheerder te maken hebt met een applicatie eigenaar per applicatie en dus per database moet er nog veel communicatie komen voor de daadwerkelijke migratie plaats kan vinden.

Op onderstaande link vind je een artikel van mij hierover op Webwereld:

Artikel SQL 2005 EOL op Webwereld.nl

 

CTP2 of the new SQL Server 2016 has been released to public

With this new beta release of SQL Server we get a better insight in what we can expect in the full product.
Some of the top capabilities in SQL Server 2016 CTP2 are:

  • Real-time Operational Analytics & In-Memory OLTP, enhanced for up to 30x faster transactions for a greater number of applications, customers can configure the in-memory columnstore to work on top of a transactional database to achieve real-time operational analytics with breakthrough OLTP performance.
  • Always Encrypted helps protect data at rest, in motion and while in use, on-premises and in the cloud.
    With Always Encrypted, SQL Server can perform operations on encrypted data.
    Best of all the encryption key resides with the application in the customer’s trusted environment.
  • Stretch Database technology keeps historical data at users’ fingertips by transparently stretching warm and cold data in a more secure manner to Microsoft Azure on demand without application changes.

I advise you to take a special look into the stretched databases.

It is a technique originating in Azure where it’s already very easy to scale up or down on a working database.
And not only on a single database!
Also a set of databases (called Elastic Pool) can be tailored to meet the per time different demand of power for each individual database.
A brand new concept which you could use in use-cases like having a database per customer or project which scale needs are different in time and in database.

In my next blogpost I will dig deeper on the new concept of elastic database.

Harry

Azure SQL Database almost on par with the on-premise version

Hi blogreaders J

I started the blog with the title: “Azure SQL Database almost on par with the on-premise version”.

And looking from a T-SQL perspective it is.
Our latest version of Azure SQL Database (let’s call it SQL Azure for now) is for 95% compatible with SQL 2014.
So not a bad title!

But there is more to it…

There are two perspectives to take in account here:

  1. Microsoft is not only working on T-SQL compatibility, but is working on al fronts to enhance functionality.
    So actually… the engine powering SQL Azure is already beyond SQL 2014 and more towards SQL 2016.
  2. SQL Azure is a PAAS offering in Azure.
    So… you don’t only get the database, but also services around it in maintenance like:
    1. Automatic replication of your date 3 times (and 3 more times if geo-redundant)
    2. Point in time restore up to 14 days
    3. Patching, security updates

Thinking of SQL Azure of being on par with SQL 2014 is kind of old skool thinking.
Thinking of SQL Azure should be about thinking about using a relational database as a service.
And those services will enhance in time.

So no more thinking about versions, but thinking about functions.
This changes the paradigm of developing big time!
You will have to develop is a much more pure way.
Keeping your application architecture clean and crisp.

For example: No more business logic in database triggers, but logic in webservices and the call’s to the database in as plain as ANSI-SQL as possible.
Code examples can be found at: https://msdn.microsoft.com/en-us/library/azure/ee621787.aspx
And to keep up to speed on the latest developments of SQL Azure keep checking the blog: http://azure.microsoft.com/blog/

Happy programming!

Link to a one-slide Powerpoint to show the latest services in Azure

Use below link to download a very usefull slide which shows you all IAAS and PAAS services within Azure:

Overall services in Azure

Have fun,

Harry

Platform as a Service (PAAS) moving a rocket speed!

As an Enterprise Architect in an organization life has always been dynamic to say the least! It is your responsibility to keep up with the latest developments in ICT both in technique as in architecture. In the old days of on-premise only that was a big challenge. But with the Cloud as a integral part of your information systems it became even more complex.

But still… The Cloud was moving vm’s to Amazon or Microsoft. So architecturally not that complex. Identity & access off course, but that’s about it. Then came Platform as a service (PAAS). That was something completely different! Not moving vm’s to the Cloud, but move complete technical workloads to the Cloud like an ESB in the Cloud, Media Services, Federated identity, Storage, etc, etc..

This does impact your architecture!

A blazing 78 new PAAS services were introduced in 2014 within Azure. So it’s moving rocket fast! And to be fair: not only at Microsoft, also are other Cloud vendors moving into the PAAS area with new services.

What is the impact for you as Enterprise Architect?

In your normal day to day work you make choices based on software you can purchase and implement at your data center. But now you should at least ask yourself for every choice you have to make: Do I want to do this myself or shall I take this as a service from one of the Cloud vendors.

An example: Your organization wants to use Cloud services from multiple Cloud vendors but you want a single sign on experience for your users. Now you can buy a federated identity server, do research on all Cloud vendors on how to connect and then build the connections. But you can also use The Windows Azure Active Directory Federation Service (ADFS) from Microsoft with over 2600 Cloud vendors already pre-installed.

Second example: You have a new web application that you need to deploy. Again you can buy a few servers, install IIS, SQL Server, the application and install everything and schedule things like backup, patch management, storage, etc., etc. But you can also take a web-role to host the web-application, Azure SQL database to host you data and let Microsoft worry about backup’s, 3 replica’s for DR, patching the server, etc.., etc.

So my message to all you Enterprise Architects out there: Examine carefully the PAAS offerings from the Cloud vendors before making expensive buy decisions. My recommendations to checkout:

Azure Service Bus, Azure Machine Learning, WAAS, BizTalk Services and Azure SQL Database. Next blog-post I will dig deeper on Azure SQL Database.

In memory technology in SQL Server

Everybody noticed the increase in technology using in memory techniques. At SAP they fully go for Hana, Oracle just started last year with in-memory database. At Microsoft we started in 2012 with in memory analytics and added OLPT in memory April 2014. The buzz is high with big marketing events, lots of whitepapers and broad press coverage.

So, but what about the real life practice?

When I visit my customers (top-50 in The Netherlands) I rarely see in memory databases used. So I always ask why they don’t make use of it. This resulted in the following reasons:

  1. I didn’t know I could run in memory with my databases.
    The marketing engine could be hitting the wrong people. Lots of database administrators are not up to speed.
  2. We don’t do it because it must be very difficult.
    True – if you use SAP then it’s common knowledge that implementing SAP Hana is not very easy. And you have to rewrite some of your programs. False – if you use Microsoft SQL Server. To start using in memory you can switch it on for certain tables (or part of tables) and without any change to the application it will work.
  3. The power of our servers is high enough. We don’t need the power.
    This is of course a compliment that our SQL Servers run so smoothly (-:

But still I think that by using in memory technology you can achieve the following:

  • Prevent hardware refresh. If servers run out of performance, moving to in memory the speeds increases again by 5x – 10x. Thus the servers can remain the same.
  • Run more VM’s on a host. By using in memory technology the number of cores can be less because of the more economic processing in memory. Thus more core’s for new VM’s on the same host.
  • Increase processing to reduce wait time for your users

So my advice: Start experimenting with the technology and look for those business cases.

My ask: anyone who has experience with the practical implementation: please reply with your live experience!

Maiden blog post

Hi all,

Today I joined my colleague Jeroen in this Blogsite about Data.
Where Jeroen is the expert in BI, Big Data & Data warehousing I tend to focus on applications, databases, development & integration. And off course with a main focus on the Cloud with Azure.

I will try to keep you posted on all new developments with my personal view on it combined with my 25 years+ experience and day to day work at my customers.
So keep posted for my new posts!

Greets, Harry

%d bloggers like this: