It only feels like a few months ago that I was writing this about HANA SPS07! In a blink of an eye, we get more HANA innovation. SPS08 was released on Friday 31st May and at Bluefin we already have our first two customer projects on it, and a third in the process of moving to it. Why? Read on.

SAP released the Business Suite for SAP HANA, and customers have been cautious. After all, most customers run their most mission critical business processes on the Business Suite – those that will cause serious financial damage to the company if they stop working. So whatever the benefit might be, customers won’t move the Business Suite until they know it’s rock solid.

That’s what HANA SPS08 is all about. The new feature list is shorter than before, but in return, the focus has been on support for the most mission critical scenarios. Tooling has been improved, there are support for more platforms and the experience is better. But, the features of the core HANA DB remain broadly similar. It’s not to say that HANA wasn’t stable before, but we are already seeing that HANA SPS08 has a more “German” car feel about it.

Enterprise Readiness

There are some nice new Enterprise features. There are more partners for Storage Replication DR, more features in the application-level DR including a new “full sync” option and better cluster management, especially from HANA Studio. It’s now possible to have multi-tier metro clusters and hostname management. In short, it’s more of the same good stuff.

We also now have support for RedHat Linux and VMWare in production, which is a step forward. VMWare support is fairly limited today, but you can be sure that will change as customers start to adopt it. IBM POWER support for HANA has not (yet) been announced but is rumored to be in progress.

In addition, there are 20 new Outsource Providers for SAP HANA and more to come.

Developers

For developers, the improvements are very incremental, but that’s not a bad thing. The Development Perspective is easier to configure, has better code completion and more wizards. Same for the Web Editor – you get recursive deletion, more code completion, better editors, function evaluation and change tracking. An incrementally improved user experience.

There are a handful of SQLScript features – extended Scalar UDF support. This is an area where HANA is some way behind other stored procedure languages like PL/SQL and this will need to be a focus in SPS09, if SAP wants to lure Oracle customers with complex apps onto HANA. There is now support for ADO.Net in the Sap.Data.Hana namespace so you can write .Net apps on HANA.

It was rumored that the Mac OS X Studio comes in SPS08 but I haven’t seen it yet. HANA Developer Edition is also now available on the Microsoft Azure cloud as well as a bunch of other cloud providers like AWS.

SAP HANA Answers has now been released as a plugin to SAP HANA Studio. This allows contextual awareness of what you are doing, your problems, and search/answer ability on how to fix the problems. This is a very cool social development tool and I worked with the product team to give feedback earlier in the year.

SAP HANA Interactive Education (SHINE) has been extended with more demos, better apps and models including a business rules model and Fiori Launchpad integration.

Modeling

The modeler got an extensive tune-up with a lot of small new features and usability enhancements. These are detailed in the following link. The highlights are usability improvements like mass copy, a new tree-map for unions, a performance analysis mode, ordering in parent/child hierarchies and better datatype management especially for currency conversation.

There is a production alert system, new color palettes and icons and better tool tips as well as error handling and integration with HANA Answers.

Unfortunately there are still missing features – especially the ability to push COUNT DISTINCT models down into the database without using code. Hopefully this will be a big focus for HANA SPS09.

The Plan Visualizer now has functionality to view physical table access plans and inverted index access display.

Monitoring, Security and Lifecycle Management

There is a very nice looking SAPUI5 dashboard for monitoring HANA now, better emergency support diagnosis, better diagnosis for blocking situations and improved Runtime dump creation – all of which will be appreciated by those of us that use HANA on a day-day basis. The resource monitor has been improved and it is possible to limit memory usage by one statement using statement_memory_limit. It’s possible to access the Plan Visualizer from the performance tab for SQL which has been monitored, which is very neat.

There are a large number of new monitoring features – it looks like since the core DB was left alone, all of the monitoring features were able to mature significantly. Check here for more details.

There are almost no new security features – just a handful of changes to auditing, data volume encryption (log volumes are still not encrypted), and the ability to create restricted users with no access at all by default. User Names can also now be in Unicode Characters.

For system lifecycle management, the unified installer is now deprecated (though I tested it and it still works in Revision 80). You must now use HANA Lifecycle Manager (hdblcm), which works for all scenarios now including rename, scale-out, etc.

For application lifecycle management, there are new features around integrating delivery units into products and packages to delivery units. The HANA Lifecycle Manager (HALM or hdbalm) has its SPS07 SAPUI5 interface improved a little. There is now change recording like in ABAP systems, that allow for versioning and transport release. There is now also a native transport manager in HANA, and CTS+ is not required. If you do use CTS+, HALM is now integrated fully into CTS+.

Predictive, Geospatial and Smart Data Access

There are new predictive algorithms for Distribution Fit, Cumulative Distribution Function, Quantile Function, Random Distribution Sampling, ARIMA, FP-Growth, CART, and K-Medoid Clustering and various other improvements. No doubt these have been driven by the KXEN acquisition, and the field has been heavily influencing which algorithms are important.

The geospatial interface gets Esri Query Layer Integration and some SHINE content, which is much needed. The former is required for Esri ArcGIS integration via ODBC, which will allow Esri to integrate fully with HANA.

Smart Data Access gets support for IBM DB2 10.1 and Netezza 7. Virtual and Physical tables can be combined in Information Views in the HANA Modeler perspective. There are various new features like being able to change a remote source, and enhancements for virtual statistics and support for VARBINARY.

Final Words

I made a bunch of predictions at the beginning of the year, and 4/10 of those have come true so far! However we haven’t seen SAP River, the HANA Graph Engine, or Complex Event Processing in HANA yet. This is aligned to the SPS08 strategy which is to focus on a mission critical core and extend the tools around the core.

For SPS09 I hope that SAP will focus on getting the new functionality that is in labs out the door and make the rest of my predictions come true. However, for me, the last 6 months have been well spent and HANA SPS08 feels like a more mature, simpler, more consistent product. I believe that SAP customers will thank SAP for this, and will repay them in a faster adoption curve for HANA.

Links to more details

I stole this from Michael Eacrett, who wrote a great blog on this subject.

What’s New in the SAP HANA Platform (Release Notes)

Web-based Development Workbench

SAP HANA Interactive Education (SHINE)

SAP HANA Platform Lifecycle Management

SAP HANA Answers

Smart Data Access

Client Interfaces Predictive Analysis Library Administration & Monitoring

Geospatial Processing

SQLScript

SAP HANA Studio Development Perspective

Scale-Out, High Availability and Disaster Recovery

SAP HANA Modeling

Security
SAP HANA Application Lifecycle Management

And stole this from Scott Feldman. Thanks Scott!

Interested in joining a Live Expert Session where a Product Expert answers all your questions you may have? Simply click and dial-in with the Participant Passcode: 1418481919 (dial-in phone numbers for all countries are attached). Each session will be recorded as well (see under Details – “Watch recorded session”) in case you missed one. You can find the session schedule below:

Date

Time (EST)

Topic

Presenter /

Product Manager

Details

June 10th

11:00 AM – 12:00 PM

Exclusive preview of SAP HANA SPS 08

Michael Eacrett

Dan Lahl

Join this session to hear an exclusive preview of SAP HANA SPS 08.

June 11th

11:00 AM – 12:00 PM

Admin/ Monitoring/ Troubleshooting

Melanie Handreck

Join this session to hear an overview of the new features in SAP HANA Administration, Monitoring and Troubleshooting for SPS 08.

June 12th

11:00 AM –

12:00 PM

Application Lifecycle Management Ronald Silberstein SAP HANA has extensive features for managing content (i.e. objects or development artifacts), transporting as well. This session provides an overview of these capabilities and discusses what’s new for SPS 08.
June 16th

11:00 AM –

11:30 AM

Predictive Analysis Library Mark Hourani Learn how to get context-sensitive help from SAP and the crowd directly from HANA Studio.
June 17th

11:00 AM –

12:00 PM

Scale-Out, High Availability and Disaster Recovery Ralf Czekalla John this session to get an update about new features mostly in SAP HANA System Replication.

June 24th

11:00 AM – 12:00 PM

SAP HANA Answers

Mark Hourani

Learn how to get context-sensitive help from SAP and the crowd directly from HANA Studio.

June 25th

11:00 AM – 11:30 AM

SAP HANA Modeling

Christoph Morgen

This session will provide an overview on the SAP HANA Modeling enhancements and new capabilities available with SAP HANA SPS 08.


To report this post you need to login first.

30 Comments

You must be Logged on to comment or reply to a post.

  1. Gregory Misiorek

    Hi John,

    eagerly awaiting POWER for HANA and more RAM for developers, i will welcome the download updates in the usual space.

    Thank you,

    greg

    (0) 
    1. John Appleby Post author

      For me, POWER on HANA is really just relevant to a small number of large customers who have standardized on POWER. Most of the largest companies in the world now run HANA, and run it on Linux/x86.

      It brings some choice, but it will always be a ported platform. Do you remember Windows NT on Alpha?

      Downloads are already up there!

      (0) 
  2. Fernando Bรกrtolo

    Cool stuff.

    Feels more mature in practically everything.

    River 2.0 is included on the download package, maybe it will be GA in SP9?

    Also, I was snooping around on the hardware checking script and found Windows references:

    http://i.imgur.com/g4nCgYi.jpg

    You heard any rumors about a Windows version? It seems there has been some fundamental rework in this version, new linux package requirements and all, maybe in preparation for future Windows support? Wishful thinking.

    Also weird that I didn’t have to edit the script to install at my home VM, using VMWare. Maybe a consequence of official VMWare support?

    (0) 
    1. John Appleby Post author

      River was mooted to go GA in SPS08, so I am confident it will hit SPS09 if not before.

      There has always been a Windows version of HANA, it was just never released. I suspect this was a slip on someone’s part ๐Ÿ™‚

      I don’t think so – I suspect there is a bug in the hardware script. Maybe Michael Eacrett can comment.

      (0) 
  3. Kumar Mayuresh

    Hi John,

    Thank you for putting it all together. SPS 08 looks more stable now, as expected. Two things I liked are: Microsoft Azure coming-in as provider and addition of handful algorithm.

    Though I am surprised that there is nothing about River as it was scheduled. 

    Regards

    Kumar ๐Ÿ™‚

    (0) 
  4. Richard Bremer

    Hi John,

    thanks for a great review and summary of the changes in SPS 8. I’ve just a little correction regarding the system lifecycle management tools. And since there has been quite a bit of to and fro in the past, I’ll try to summarize the situation:

    The one and only system lifecycle management tool (sometimes SAP might say platform lifecycle management tool) is hdblcm (not hdbalm), which has a gui variant named hdblcmgui. Both are part of the installation medium, the component downloads, and also residently installed with each database system.

    For anyone who has used any of the other “system lifecycle management tools” like the HANA Lifecycle Manager HLM or the unified installer, hdblcm must be like Christmas and the announcement of finally no more star wars movies on the same day – that’s my personal opinion. Most of the lcm tools (hdbreg, hdbaddhost, hdbremovehost, etc.) have in SPS 8 been integrated into hdblcm(gui), so that you can now manage all platform lifecycle aspects of the system from one single tool.

    Unified installer has now been deprecated (it was still shipped with SPS 7 but usage discouraged). And in the HLM tool, all platform lifecycle management functionality that is covered by hdblcm(gui) has been disabled. HLM used a different backend implementation for these tasks than hdblcm. I expect that HLM will be moved to use the backend implementation of hdblcm and then these functionalities might again be available in HLM for those admins who prefer web-based access, avoiding ssh logons.

    Finally, hdbalm covers application lifecycle management such as deploying applications, delivery units, and so on. It’s the command line tool that can be used for batch operations – more or less all functionality is also contained in the HANA Application Lifecycle Management application HALM (http://<host>:<port>/sap/hana/xs/lm/).

    While hdblcm must naturally be running locally on the database server, hdbalm is, as you said, also available on client hosts.

    Best,

    Richard

    (0) 
    1. John Appleby Post author

      Thanks Richard. I’ve corrected this, I think.

      By the way the unified installer is still present in SPS08 and does still work.

      Also by the way we tested hdblcm for SID rename in Rev.74 and it broke the HANA storage and the HANA system needed a rebuild. Hope this has now been fixed in Rev.80 though we won’t be testing it ๐Ÿ™‚

      (0) 
    1. John Appleby Post author

      Technically the Extended Storage came with SP07, at least in pilot release. I haven’t seen this status change in SP08.

      HRF – great spot, I hadn’t noticed this made it in. I worked a little with that team earlier this year and it looked super-interesting. Gary Elliott will like this a lot.

      RedHat, nice spot on the announcement on SAP Service Marketplace, I’ve linked it here.

      (0) 
      1. Gary Elliott

        Thanks Jon, I’ve been looking forward to HRF, as you know. Can already think of a number of customers where this would be very useful.

        (0) 
  5. Mikhail Budilov

    John, thank you for review,

    HANA still is not protected from memory overflow.

    You can still write select * from billion_rows_table and get OOM.

    Global parameter statement_memory_limit is good idea, but it’s not a silver bullet, need more steps. indexserver must not be restarted because of OOM or SDA, OOMs must be in the past.

    (0) 
    1. Justin Molenaur

      Yes, I was under the impression that SPS8 would include some type of limits, whether per statement per user or per session. I can’t recall where the conversation took place, but I also see this as a very strong need as runaway queries can still take the system down.

      Regards,

      Justin

      (0) 
        1. John Appleby Post author

          The 2bn limit is a design choice/limitation, because it uses a 32-bit integer for dictionary encoding. This helps with compression. Better to use partitions than move to a 64-bit integer. I’m pretty happy with that choice myself.

          For sure HANA needs better workload management, but I see other more pressing development priorities. I won’t air my laundry list in a public forum though)

          (0) 
          1. Mikhail Budilov

            In IQ – it’s also columnar db with bits dictionary encoding – haven’t any billions limits.

            Number of rows per table

            Limited by table size, upper limit 2^48 -1

            Physical Limitations – Audience – SAP Library

            HANA Limits Restrictions for SQL Statements – SAP HANA SQL and System Views Reference – SAP Library

            Vertica limits – vertica also columnar db

            https://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/SQLReferenceManual/SystemLimits.htm%3FTocPath%3DSQL%20Reference%20Manual|_____2

            (0) 
            1. John Appleby Post author

              Right and the 32-bit dictionary encoding design decision means that HANA is faster than IQ because it can use SSE2 instructions to do multiple instructions in one cycle. You see this in my benchmarks

              It’s a design trade-off. HANA is full of design trade-offs that way.

              (0) 
                1. John Appleby Post author

                  Well yes but more than this, the Intel x86 platform is designed to be able to very efficiently process 32-bit. For instance the instructions PADDD PSUBD etc. operate against 4×32-bit integers in one instruction on one core.

                  If HANA were to use a 48-bit dictionary encoding then a number of functions would need to be 64-bit integers and would be 2x as slow.

                  From IQ’s perspective this isn’t critical because IQ supports multiple database platforms, so it can’t make those sorts of design assumptions.

                  But it explains why HANA has a 2bn row limit per partition, and also why HANA is faster for certain operations.

                  (0) 
  6. John Studdert

    Great blog post John, thanks for the info. One quick note: the dial-in phone number list link for the Live Expert Sessions gives me an “Unauthorized” error.

    (0) 

Leave a Reply