Friday, January 24, 2014

What is "the cloud"? - I think yesterday it became a bit more clear

Yesterday, i.e. January 23rd 2014, I attended the IDC event "360 degrees of IT" in Copenhagen. It was a good event with speakers from IBM, Symantec, Hitachi, T-Systems and others. The focus of this years "360 degrees of IT" was what IDC call the 3rd platform, i.e. mobility, big data, social and cloud.

But what is the cloud, that all IT-person keeps talking about. In the past I have learned, that the cloud can have either a private cloud or a public cloud. But what is this cloud? During today's event it became a little clearer during the talk by Tony Franck "Increased business value with cloud". He talked about Hitachi's Unified Compute Platform (UPC). According to Franck this UCP makes it possible to deploy new servers in just 46 minutes by inserting a couple of blades. Blades? So, it appears a cloud, at least the Hitachi UCP one, involve blade servers. So I guess a cloud is just a sophisticated box with blade server and storage well integrated. Much like Oracle's Exadata or Exalogic boxes or IBM's System z mainframes. Only difference is, that on a System z mainframe no hardware need to be added to deploy new servers and the process only takes a few minutes and not almost an hour (unfortunately I don't have any information about the similar processes on Oracle equipment. Really I see little functional difference between an Exalogic box and a System z mainframe - except the name on the box. Am I seeing the world of the cloud as it really is? I.e. old wine on new bottles.

Arne Sigurd Rognan Nielsen, who is a Norwegian that have worked many years with IBM and whom I have heard before, came and talked about Social Business 2.0. Not as a service offering from IBM but as a process. Among the questions he asked were: "What is the business value of a license?", with the answer "Zero". I guess the question was asked from the point of view of the customer - not the vendor. Unfortunately Arne's Apple laptop did not play well with the projector in IDA's main conference room, so we could only see every second slide during the presentation.

Another presentation, which deserves to be mentioned was that of T-Systems' Dieter Weisshaar "Life and Business Changes - Zero Distance" in which he showed a grocery shop at a Japanese subway station, that just consisted of pictures of groceries with QR-codes on. The customers simply scan the QR-codes of the goods she wants to buy, and then they are payed for using the phone, and later that day delivered to the buyers home. I consider that the first intelligent usage of QR-code.

Saxobank's Mikael Munck talked about using the majority of his IT budget on new developments year after year, and about a new social trading platform announced by the bank the same morning called TradingFloor.com. Part of secret of spending most of the budget on new development is focus on just one platform, i.e. .NET, and the needs of the traders, i.e. the customers. Mikael Müller from DR Technology talked about the transformation of the broadcaster from the old static world in which they decided when, where and how to the new dynamic media consumption model, where the user/customers decides when, where and how. Examples mentioned was recent coverage of municipal elections in Denmark from 56 locations using mobile phone video technology, and harvesting tweets in almost real time as part of the coverage of this election.

I found that even the I am not currently in the market for any of the services offered by the companies taking part in "360 degrees of IT" the day was time well spent to get updated on current trends. Although I was a by surprised to see to presentations with copyright notices from respectively 2012 and 2011 at the end.


Sunday, January 12, 2014

It used to be my favorite periodical!

Many years ago, while I was chemical engineering student at the Technical University of Denmark I remember reading Hydrocarbon Processing regularly. I seemed to me, that the issues at that time was significantly thicker then the current monthly issue. I always find it fascinating to read about the many large construction project going on across the globe.

Lately I have been somewhat disappointed with the articles I have chosen to read in Hydrocarbon Processing. My primary field of interest is process safety so when received the December issue some weeks ago I was pleased to find to safety related articles: "Rethink the hazards in your process" by R. Modi and "Consider process-based failure analysis methods for piping and equipment" by D.L.N. Cypriano et.al.

I started reading the first of these and after reading the introductory paragraph I started wondering about the authors purpose of writing this article. It indicated, that process safety was just about "meeting mandatory requirements". I was a bit shocked, but continued reading.

The next three sections were titled respectively "Managing process hazards safety", "Safety life cycle" and "Basis of SIS design". This appeared interesting, although I was wondering about the role of the word "safety" in the title of the first section.

It turned out that the section "Managing process hazards" was mostly about designing and operating the process within acceptable risk limits. I was stunned to read, that "eliminating risk and still operating the facility in absolute safety is not practically possible".

The section "Safety life cycle" starts with mentioning the functional safety standards for electronic safety systems IEC 61511 and IEC 61508. But the author states that "to balance system design between safety and availability". Does the author really mean, that increased process safety means reduced system availability? I hope not! The author also seem to believe the analysis and realization phases of the process life cycle are more important than the operation phase. I respectfully disagree.

The section continue with stating, that the major to comply with IEC 61511 are SIF identification, SIF assessment, SIF design and SIF validation. However, when the author goes on to write "Risk assessment determines the required safety instrumented level (SIL)", then it leaves me with the impression, that the knowledge of IEC 61511 could be deeper. I also wonder why any discussion of the importance of company's allowable risk criteria, which are required to determine a safety integrity level, appear to be missing.

The physical components - input elements, logic solvers and output elements - of a safety instrumented system are correctly identified in the section "Basis of SIS design". But the sentence "...compenents are assessed for their complexity, inherent properties and behavioral uncertainty" I find very difficult to understand, and the same is the case with the sudden introduction of type A or B subsystems immediately following.

Unfortunately a direct error appear in Table 1, which list factors to be considered in SIS design. Here the third factor is "Keep PFDavg value of each SIF superior or greater than the targeted value" should be "lower than the targeted value". I am not an expert on the design of safety systems - that is why I read this article - and maybe that is why I find Figure 3 difficult to understand. Especially that PFDavg = SIL is a contributing factor for SIS voting. Maybe the word "voting" in Figure 3 should have been "design"?

Some times things are explained with a few more word than really needed, as in "..., PFDavg, as the probability of failure to perform the desired safety function when demanded". Most engineers would just write "..., PFDavg, the probability of failure on demand".

The following section titled "Contributing factors for SIS voting" first list all the factors / calculations, which enter into a SIS design, and the continues the discussion about safety versus availability without in my view making this so-called "balance" any clearer. However, I get the impression, that SIS voting is just choosing which XooY voting structure to implement. Can it be that simple?

The final sections are titled "Other factors", "Effective factors to design an SIS" and "Other measures for SIS life cycle design". The first of these continue the discussion of voting structure after again listing a number of factors, which should be considered in SIS design. As far as I can tell the section "Effective factors to design an SIS" don't contain a single word relating to effective design as least from an engineering perspective. Final section "Other measures for SIS life cycle design" state the important and surprising fact, that "the optimum is the best design".

My conclusion is that the title of the article has very little relation to its content, and hence I recommend, that in stead of wasting time reading this article, then go directly to the book "Safety Instrumented System Verification" by William Goble and Harry Cheddie mentioned in the bibliography.

I other article mentioned at the start of this blog is better. It contains a nice case study of investigation of leakages in a boiler feed water system. But again I have difficulty seeing link between title and content. I appears the editors of HP want titles to be as broad as possible to increase readership. But aren't engineers smart enough to see through that? Maybe HP articles are not aimed at engineers?