First Take - SAP's IoT Strategy becomes clearer....

Like many industry observers, I was also surprised when SAP started to prominently feature IoT as a key message by mid of 2014. Not that IoT is not an important trend and a key strategy to transform the way how enterprises and their customers deal with ‘things’ – so the revenue potential is there – the question is what can SAP really contribute to the immense technological challenges that encompass IoT. The questions started to pile up as conversation with customers almost never pointed to SAP as a solution in the space, so getting (finally) briefed by SAP on the IoT strategy was very welcome.
 

First of all it is clear that SAP is committed to the space, announcing 500 (incremental?) developers to be dedicated to IoT, not a small change numbers. And it is also good to see that SAP will use the HANA Cloud Platform (HCP), its PaaS product for the effort (more here). So let’s take a look at the first three use cases that SAP wants to tackle:

• SAP Predictive Maintenance and Service – SAP wants to help enterprises be better at maintenance and servicing ‘things’. Combined with the augmented reality (AR) capabilities for the service technician certainly a compelling scenario.
   
• SAP Connected Logistics – In this use case SAP wants to connect logistical ‘things’ (e.g. trucks, containers, railcars etc.), combine them with 3rd party data (e.g. weather, traffic etc.) and tie them with its existing logistics and transportation management solutions. Again there is an AR angle with the support of smart glass devices for warehouse pickers.
   
• SAP Connected Manufacturing – Here SAP uses its next version of Manufacturing Execution (15.0), puts it on HANA and with that allows for typical IoT data as from sensors, beacons and RFID chips to be part of the manufacturing execution puzzle.

Next let’s look at the technologies SAP has at its disposal and SAP wants to use in these IoT use cases:
 

• Hana Cloud Platform (HCP) Today’s IoT projects are inherently custom development projects built on top of components that are more or less tuned or built for the unique challenges of IoT. Modern software gets built with the help of a PaaS and SAP has a solid option with HCP. But HCP will have to add the capabilities to deploy HCP built applications easily to standard cloud platforms, AWS most prominently. Certainly an overall good feature for HCP.
   
• HANA with all things SAP these days, HANA features prominently. And HANA’s fast in memory analysis capabilities bring a certain level of attraction to IoT challenges, but fast in memory analysis is only a subset of all the ingredients needed for IoT projects. Unfortunately for SAP, for many IoT uses cases requiring fast in memory analysis capabilities are optional or nice to have. The ‘bread and butter’ of IoT are cheapest storage and unstructured retrieval databases – to cope with data volume as well as unforeseeable analysis questions.

SAP Response - HANA has addressed this concern with Dynamic Tiering, using a columnar disk technology integrated with HANA that supports unstructured data storage and retrieval (and is integrated with the textual capabilities of HANA in-memory).  This address the cheaper storage and analysis of unstructured data and is in market with HANA SPS09. In addition, HANA has integrated the Sybase ESP streaming data analytics (CEP) engine into core HANA functionality (HANA Smart Data Streaming), so now lightweight analysis and correlations from streams like sensor data can be accomplished in HANA Smart Data Streaming, and additional data can be passed from Smart Data Streaming to HANA for deeper realtime analysis.  With Smart Data Streaming, streams of data (filtered or unfiltered) can be passed to multiple tiers of data for storage and analysis, including SAP  HANA, SAP IQ, and Hadoop.    

My response -  Fair enough to raise that unstructured data analysis can be done on cheaper storage with dynamic tiering, certainly a key feature in HANA SPS 09. But the queries will run separate, not on the same data and need to be combined, which makes the engineering process more complex (something I am sure SAP will take care of) - but may ultimately exclude customers from some insights. Running two separate queries and combining the results may but must not lead to the same result as running one query on the combined data. Material for a separate blog post. 
And fair to point out the SAP has used the Sybase ESP / CEP capabilities and added them to the HANA core, always good and great to see re-use of working and proven software. CEP certainly has its space in IoT applications. 

• SQL Anywhere The venerable and proven SQL Anywhere is a viable player for IoT. Not all ‘things’ will be connected to the internet, for a number of power, health and practical network coverage reasons. The ability to store vast amounts of data reliably, low cost on a HDD or SSD chip over prolonged times is a must have for a number of ‘things’. The Sybase built replication capabilities at a time of connection are an extra bonus.
   
• Sybase Data Management With Sybase SAP has acquired also a sizeable number and decently capable number of data management tools. And shuffling data around is an inherent nature and characteristic of IoT projects. The Sybase tools are good for e.g. moving data from Hadoop or Amazon S3 storage into e.g. HANA or Sybase IQ. The question of course to be addressed – why not process data where the information occurs in the first place.
   
• Business Applications There can be no question SAP has the relevant business applications in CRM, SCM and PLM that are essential for IoT. The key area to watch will be though, if the gravity of these applications can pull the IoT data ‘in’ or if it is more likely that a subset of that ERP data will be moved to the ‘rest of the IoT’ data for in situ analysis where most of the data is. Too early to see a trend but a key area to watch. 

In summary SAP certainly brings some interesting and valuable scenarios for IoT in the first wave of offerings. On the technology side SAP also brings some interesting products to the table, but we will have to see how they all will play together and how they can enable the next set of customer use cases in the area of IoT.

MyPOV

Learning more about a vendor’s offering is always valuable and creates a better understanding for the angle the vendor has. In SAP’s situation the angle is the one of the business application vendor with a very fast analytical in memory database. And there are certainly a number of challenging IoT scenarios that require real time capabilities, closely intertwined with business applications. Think off real time manufacturing, mass customization, individualized service and real world traffic situations. Not really SAP’s core competence (so far), but certainly scenarios that command a premium price in the IoT world, something SAP certainly wants (and has to) zero in on.

But if this slice of the IoT market is large enough for SAP in the medium run, is something that remains to be seen. Entering the IoT market with none direct offering for Hadoop style BigData and HDD / SDD based cloud storage excludes SAP from the most common IoT projects we see with customers. But what isn’t possible today – can happen tomorrow. Rest assured we will be watching. 

• SAP appoints a CTO - some musings - read here
• Event Report - SAP's SAPtd - (Finally) more talk on PaaS, good progress and aligning with IBM and Oracle - read here
• News Analysis - SAP and IBM partner for cloud success - good news - read here
• Market Move - SAP strikes again - this time it is Concur and the spend into spend management - read here
• Event Report - SAP SuccessFactors picks up speed - but there remains work to be done - read here
• First Take - SAP SuccessFactors SuccessConnect - Top 3 Takeaways Day 1 Keynote - read here.
• Event Report - Sapphire - SAP finds its (unique) path to cloud - read here
• What I would like SAP to address this Sapphire - read here
• News Analysis - SAP becomes more about applications - again - read here
• Market Move - SAP acquires Fieldglass - off to the contingent workforce - early move or reaction? Read here.
• SAP's startup program keep rolling – read here.
• Why SAP acquired KXEN? Getting serious about Analytics – read here.
• SAP steamlines organization further – the Danes are leaving – read here.
• Reading between the lines… SAP Q2 Earnings – cloudy with potential structural changes – read here.
• SAP wants to be a technology company, really – read here
• Why SAP acquired hybris software – read here.
• SAP gets serious about the cloud – organizationally – read here.
• Taking stock – what SAP answered and it didn’t answer this Sapphire [2013] – read here.
• Act III & Final Day – A tale of two conference – Sapphire & SuiteWorld13 – read here.
• The middle day – 2 keynotes and press releases – Sapphire & SuiteWorld – read here.
• A tale of 2 keynotes and press releases – Sapphire & SuiteWorld – read here.
• What I would like SAP to address this Sapphire – read here.
• Why 3rd party maintenance is key to SAP’s and Oracle’s success – read here.
• Why SAP acquired Camillion – read here.
• Why SAP acquired SmartOps – read here.
• Next in your mall – SAP and Oracle? Read here.