Thoughts on SapphireNOW ’14!

I watched Bill McDermott’s first keynote as CEO of SAP. His decision to begin the keynote with 3 students presenting their thoughts was a bold move . Unheard of in SAP history. I loved it. I loved it not because it was a bold move but because I’ve a story to share.

Lean & Hungry Look (LHL)

In July 2011, my son & I attended SAP InnoJam at Palo Alto, Ca. My son then was rising sophomore in an US University. We had a great experience. Lovely memories. The SAP team(headed by Anne Hardy, Juergen Schmerder & others) who organized was fabulous. That event provided my son with an opportunity to experience the corporate environment for 24+ hours. Soon after we landed in San Francisco from Atlanta, my son fell in love with Silicon Valley. He wanted to move there after graduation. Early next morning we attended InnoJam. Met lovely people. His love for Silicon Valley rose exponentially with InnoJam experience.

He was planning to specialize in Bio-Medical Engineering in freshman year. He spent his summer interning at an IT company in North Carolina. He attended InnoJam while interning there. After completing the internship & InnoJam, he decided to specialize in Computer Science.

He finished next two summer internships in Silicon Valley. He would be beginning his career as an APM(Associate Product Manager) in San Francisco, CA this August.

When I began my career in ’80’s, the CEO used to tell one of Shakespeare quotes to us employees: You all should look lean & hungry. 3 kids Bill McDermott/Hannah Datz presented to us in SapphireNOW looked lean & hungry. With that look comes impatience, not little but a lot of impatience – a quality which is suicidal in big corporations such as SAP. That CEO also used to hate “*ing” as in trying. This continuous tense will not take us anywhere, ing provides people with a platform to procrastinate, he used to say. My son unfortunately didn’t show an interest with SAP technologies even though 3 years ago we had a great time with SAP. I can only speculate the reasons why he was not interested: SAP didn’t make their presence felt in college campuses. Plus it took several months – SAP used *ing liberally – to get trial version of SAPSUP(Sybase Unwired Platform). We used this product in InnoJam but couldn’t continue our work. Had we continued our work, my son probably wouldn’t have forgotten SAP. I know Bill McDermott has a desire to make a difference & fix this issue.  This is catch-22 situation. SAP needs people who’re impatient. SAP’s culture require people to be patient. Good luck.

Mission Driven

Hannah Datz mentioned her generation is mission driven meaning they’re interested in giving back & help society than getting a big paycheck. SAP needs them, needs lot of them. Mr McDermott has taken initiative to influence the younger generation to join SAP. Not just by words but by actions, presenting them to SapphireNOW audience. Please continue. As dad of 22 year old son, I can tell this: SAP didn’t focus in hiring LHL & mission driven people in the past. Change that please. I hope he would get all support he needs within SAP to promote that practice.


The message I heard in SapphireNOW was not simple. I know nothing changes overnight but at the same time, we’ve heard “*ing” thing several times in the past – fixing issues around certification, education & training, trial versions to name a few. Marketing two polar opposites products SFIN & BW don’t make our(independent consultants, customers etc) job easier & fun. The SFIN message last week was not simple. I heard – from people who attended – 3 different versions of what SFIN is in relation to SoH. Finally I read this blog: It provides some details on SAP’s direction. I’m still not clear on roadmap to implement HANA. We’ve a lot of options but not with a lot of details or pros & cons for each. Hope we’ll start seeing more clarity in days/weeks/months to come.

My best wishes to SAP.

Oracle 12c – Two Copies

I watched OOW13 keynote y’day. Oracle CEO Larry Ellison mentioned Oracle 12c would maintain two copies of data, one for OLTP & another for OLAP. And data for both OLTP & OLAP tables would be updated synchronously. OLTP tables would be stored in row format & OLAP tables would be in Columnar format.

The columnar format OLAP tables would be helpful in reading data faster while row format OLTP tables would help in processing OLTP transactions faster.

It sounded great soon after hearing. On further analysis & discussions, I’ve following questions:

1) Mr Ellison mentioned OLAP queries would run faster as data would be in the columnar format. I get this. He also mentioned OLTP transactions would run faster. The reason: Current systems have 10-20 analytical indexes per table & 1-3 OLTP indexes per table. Since users are going to use OLAP system for queries/analytical requirements, we would no longer need to create 10-20 analytical indexes; as a result, updates/deletes/inserts would perform less number of operations. This would result in improved performance he reasoned. I would agree with him if all tables have 10-20 analytical indexes on an average.

However, I don’t believe I’ve ever worked in a system(including ERP system such as SAP) which had/has 11-23 indexes per table on an average. At the maximum, I’ve seen a few tables having ~10 indexes(total). On an average, I’ve seen 3-5 indexes per table. And as customers normally have some kind of Data Warehouse systems for their analytical requirements, almost all indexes are used for OLTP purposes. In that case, I doubt a significant number of indexes can be dropped to improve the performance. 10-20 analytical indexes seem to be a very big assumption. If this assumption is not correct(not by a little but by a big margin), then wouldn’t we see a significant OLTP performance degradation?

2) How would this approach work with SAP? In order to improve the scalability & performance of SAP systems, SAP follows asynchronous processing using what’s known as VB* tables. In other words, SAP uses their own transaction handling. SAP doesn’t rely entirely on Oracle’s transaction handling to ensure the data consistency. Would Oracle use VB* tables to copy data to OLAP side or trigger updates/inserts/deletes etc only when target tables are updated asynchronously? How would the error on OLAP side be handled if processed asynchronously?

3) SAP also uses the application servers buffering. How would this be updated to OLAP side? Wouldn’t there be an inconsistency between OLAP side & OLTP side because SAP knows a little bit more about data(based on whether table is buffered or not) than Oracle?

4) What are memory/disk size requirements to manage two copies of data?

5) No logging for OLAP data. What does this mean to recover the database from a crash? Would all OLAP data be recreated during restart after a crash? Would that mean the recovery time after a crash be a lot longer?

Availability of Reasonably Priced Expert Resources!

I was reading this blog.

Dennis Moore says:

However, in conversations with SAP HANA customers and experts, there are still significant issues to be overcome before this becomes a worthy competitor to Oracle – issues in cost, stability, scalability, standards support, openness to a third party ecosystem, availability of complementary solutions, and availability of reasonably priced expert resources.

I wonder if non-availability of reasonably priced expert resources could be the primary reason for all other issues(except cost) listed by Dennis Moore. I say this because the expert resources would’ve identified other issues(assuming they’re true) quickly. Quicker identification would’ve led to quicker resolution. I don’t know what the author thinks as the reasonable price. As far as I know, SAP-HANA market is not attractive for the independent, detailed, ambitious, determined, dedicated, hard-working, successful consultants with several years of SQL / RDBMS / Unix / Linux / SAP / scripting / programming / data modeling / DW / DBA experience but with little or no HANA experience.

I was reading another blog  by Vijay. In this blog, he tries to define what big data is. Based on reviewing all responses of the customers, he concludes :

There is only one big V that matters to customers – BIG VALUE.

Now let us assume non-availability of reasonably priced expert resources is an issue. Let us assume SAP HANA marketing team is one of SAP customers. Let us also assume that identifying & training appropriate resources for HANA expertise would’ve provided BIG VALUE to the marketing team. What data could’ve been used by SAP to get that BIG VALUE?

  • SAP’s internal database which keep track of all messages opened by the employees & the independent consultants. This database would’ve provided information on the length of SAP experience & the quality of messages. SAP could’ve identified all consultants with n+ years of experience with additional criteria (such as experience with large corporations, length of contracts, length of non-SAP IT experience, quantity & quality of messages etc) & created a target group of consultants. Gather & analyze information for the target group from other sources such as:
  1. The level of participation in SCN – look for both the presence or absence of participation. Absence of participation may indicate busy work schedules.
  2. LinkedIn profiles & the recommendations
  3. The length & the scope of contract work performed by the consultants
  4. Other activities (either presence or absence) in social media including the presentations, webinars, tweets, InnoJams, TechEd/Sapphire participations etc. Once again the absence may indicate busy work schedules
  5. Other soft skills such as the level of aptitude demonstrated by attending the training programs, certificate exams etc.

I don’t know if SAP used the information from the databases identified above to recruit & train resources for developing HANA expertise but the point is that BIG DATA is nothing to do with the amount of data(Volume) or other Vs as confirmed by Vijay.

A Quarter Million Dollars Dilemma…

On April 4th & 5th, my daughter & I visited Penn to attend Penn Preview day for ED(Early Decision) applicants. Almost a month back, she became one of the fifty GaTech’s coveted President’s Scholars(full ride to GaTech). The difference between attending GaTech versus Penn is a little more than a quarter million dollars over the next four years. In terms of quality of education, I’m almost sure there’s no big difference between these two great institutions. However, based on personal experiences, I’m aware the college education is not just about tangible things such as a degree, a job etc. Several intangible benefits are associated with the college education.

Personally I became a better person when I went to the Regional Engineering College, Trichy (RECT)  in 1980 when I was 16. After graduation, I ended up working for a factory about 100 kms away from that great institute. I spent approximately 3 years there. Every minute I was there, I knew I didn’t belong there. I wanted to get out of there, right away. Had I gone to another school, I might have become a better engineer and would’ve loved that job. Actually other trainees who joined with me from other engineering colleges, loved that job. At RECT, I, for sure, became a better person.

As a result of my own experiences, I wanted to learn about Penn & participate in my daughter’s decision making process. Ever since she received her acceptance decision in Dec, she was almost certain about her choice: She very badly wanted to go to Penn. As a side note, she also changed her mind on Engineering. Three or four years ago, she didn’t want to become an engineer. She definitely didn’t want to study engineering. I was not sure why.

Two or three years ago, I met two great women Marilyn Pratt & Anne Hardy, both work for SAP. More details on Anne Hardy: Both work to raise the awareness of the issues around Women In Technology(WIT). I shared my thoughts learned from them with my daughter. She slowly started changing her opinion on Engineering. Along with WIT thoughts, she also started realizing she was more comfortable with Math & Science than other subjects. As a result, she decided to apply to Engineering School. When she applied, she was still not 100% sure.

We attended Penn Preview Day on April 5th. As a part of the event, we attended a luncheon meeting with an engineering student & a professor. While we were waiting at the airport for our return trip home, she mentioned she, based on the luncheon meeting, would like to definitely do engineering. My daughter’s transformation from “Definite No” to “Definite Yes” to become an engineer was just incredible. The luncheon meeting was remarkable. The junior student we met looked highly motivated & knew what she was doing. Both she & her professor were knowledgable & answered all our questions about Penn & Engineering. I don’t know if everyone who attended the event was lucky enough to meet a student & a professor of very high caliber but we did. I was impressed with not just that student but all other students as well.

On April 4th, we met another student whom we knew from his childhood. He was another kid who not only knew what he was doing but spent time with us despite his busy schedule. His thoughts about going to Penn reminded my own thoughts about my college experiences.

Overall I’m very pleased with my trip to Penn.

At this point, as my son would say, the decision is not about picking the right college; but rather what we’re willing to give up…