Passionate About Tandem
Ross Systems International Limited
Tandem/HP NonStopTM
Development & Test Solutions
|
Passionate About Tandem Ross Systems International Limited
Tandem/HP NonStopTM
Development & Test Solutions |
|
|
Mistley Telegram Archive Issue 2006-2007 November 6th, 2007 The path of progress is rarely straight or simple, and so it was last week. I started off with good intentions to get cracking on the Universal IP Listener and upgrade it to include a dynamic operator control interface and multiple IP Listeners. Printed out the source code and started crawling through it making amendments, which were quite a few, when it slowly dawned on me that I would need to backtrack in case of disaster, i.e. version control. Well, some time ago we had brought a copy of PVCS to help us support one of our clients, had installed it, played about with it a bit and then hit problems. So, since I was rather busy, I shelved it and used good old methods of version control, which work very well with small systems. However, in the last few years we have changed and developed a great number of products, with an enormous number of source files, etc. Since, I had used PVCS as a user for many years, I though what could be simpler? Ah, but with great power comes great responsibility, and the problem was that as a novice administrator, after installing the product, I had done a number of things which one should NEVER do, for instance setting up a project database in the root directory, which I later found prevents you from setting up any other project databases on that volume, a fact which I only found out after several turns of the installation handle. Having solved that problem, the rest of the week was thus spent archiving and cataloguing of all our products, with a brief pause for writing an article for TandemWorld about our new version of FINFO so that at least it's users will be warned that the test version license expires after 2007. This week, I am continuing the archiving of our products, both internal and external, yes we do do concept R and D, and hopefully get back on track with the update to our RIPPS product. I'll let you know. October 29th, 2007 I seem that interest abounds over our FINFO ordering page, but no messages, however I have also noticed that our contacts page was becoming a hot favourite. So it seems that you are considering buying FINFO, but need to put it to your management, so I've linked in the FINFO Terms and Conditions, so that you can present those as well. I look forward to hearing from you. Any problems, please contact rupert@rsi-ns.com. Otherwise, I have had discussions with some of our American friends, and they are telling me that our IP Suite RIPPS, is just what many people need when they are implementing Active-Active solutions. Tandem users have the application software, but they also need to add in many, many IP stubs to perform the communication. Whereas with RIPPS all the user has to do is to shoot the message at the IP stub ($process.#ext) and the message gets shot off over the ether, and on the other side there will be the corresponding listener to do the link to another process or pathway server. Following this interest, I am upgrading the quality of the listener so that it can listen to many ports at the same time and feed the messages to the corresponding processes, rather than a single port and process for each listener, together with the same smooth operator interface as exists with the client IP stub. I have also got to thinking that you might want this session encrypted, any interest, please let me know. So, I am going to be digging into the C++ code this week to do the job, it's not difficult but having multiple IP listeners running from the same process is a little unusual but very easy to manage because C++ gives you access to all the object handles. Back to the grindstone! There is work to be done! October 24th, 2007 Firstly, I want to make a huge apology to all of you who tried to use our FINFO ordering page and found that it did not work. This has now been corrected. Sorry. Last week was hectic as I prepared to give a presentation and demonstration of our HSM Emulator product, something which of course I have done before, but nevertheless it is stressful when you not only have to demonstrate the features of your product but also ensure that all the communications work to enable remote working on your system and at the same time to maintain a high level of security. It worked technically, but I am not so sure that I did not try to do too much, perhaps I should have been more selective about what I tried to include. I suppose it was all a learning experience. However, with the extensive checking before I did the demonstration, I now know that the product as it stands is solid and fulfils is original requirement of being easily upgradeable to any customers requirements. The HSM Emulator was always intended to be a security applications framework to speed the development of HSM based products, with the resultant improvement of ROI and the quality of the finished applications product. The rest of this week will be consolidation on suggestions made at Brighton, chasing up leads and maybe, hopefully, processing FINFO orders from our order form on our web site. I'll let you know what happens next week. October 13th, 2007 A week is a long time in politics, so they say. Politics was
never my forte, but it's been a long week. The European ITUG, otherwise know as
ETUG, took place this week in Brighton. The seminars covered a multitude of Tandem oriented subjects and were very good, with the notable exception of one vendor seminar, which appeared to be carried out by a gibbering monkey showing minute and incomprehensible slides. It lasted for 15 minutes rather than 45, thankfully. Do not ask me who it was! Such things are left to the oblivion of history and act as a salutary lesson to those involved, who will only too well remember opportunities lost. Note well, it is very easy to descend to these abysmal depths and I am only too willing to admit that I too have had my moments. The accreditation exams were also well attended by those hardy and determined techies who realised that having bits of paper to say they know what they say on their CVs is not just a load of moonshine but has been backed up by some intellectual effort as well as experience. The only problem is that the exams are much harder for developers, who normally never get to see the kit they are programming, so it's out with the manuals and oh my! does it all look like this and now I know how to get round my problems, lets try it! Crash! The computer manager has been reading "HP NonStop Server Security" by "XYPRO", and would never let his developers touch his precious infrastructure. I must admit that I too took part and managed to get accredited, if not certified. However, the manner of taking exams has altered over the last few years, so I had to re-register and only hope that HP manages to join the two ends together and send me my card. Yes you get a blue card for all your effort, as Mike Jones, who has nearly the full set showed me. The exhibitor stands were of great interest and showed how third party products can have a major impact on the ease of use and security of NonStop systems. As a result of which, besides building what I hope to be long term relationships, I have also been persuaded to try some of the products, thus TELNET communications over IP to our machine will now always be secured, although that does not protect us from key logging spyware. However that all comes under the heading of Information Security Management Systems ISMS, as one of the presenters was trying to say but never got round to it. It is sad sometimes how people get so close to the crux of the matter, pause to crack it open and then miss. There is nothing to beat a good anti-climax. The networking was fun. It is so good to see people in the flesh who you only have ever conversed with by e-Mail or telephone conversations, which because of the time difference can only take place when you are at the end of your day and they are at the start of theirs. Also, the organisers' fears for the exhibits at the Brighton Museum were thankfully unfulfilled, but the structure of the building was ideal. Because rather than having everyone in a vast space, as often happens, people could spread out throughout the museum and cluster together into small groups to chat. The result was a great success, also the wine and the food were fabulous. Thank you BITUG committee! HP were at their very best and assembled experts in every aspect of the Tandem/NonStop platform, as well as an impressive array of financial expertise for their financial forum, and I suppose that is what made ETUG Brighton such a success. It was whether your interest in the Tandem Platform was technical, financial or managerial, there were the people there who were very knowledgeable and skilled in their subject, and had years of experience, who could and would give you the answers that you always wanted to know, but could never otherwise have found out. Normal life is very ordinary by comparison, see you next week. October 7th, 2007 Long Time No See, as they say in Chinese. But that does not mean the nothing has been happening. Over the last month I have been busy working on the Licensing Manual which is not ready for its pre-release. In other words, it's almost all there and ready for those of you who mite be interested in using our Licensing product for managing access to your programs. If you are interested you can find me at ETUG in Brighton and we can have a chat. An interesting thing however about writing manuals is that it makes you think about what a user sees and the usability of your programs and systems. Hence a large number of small changes have been put into the licensing system to make it more user friendly. The other activity which I have been engaging in is the release ready version of FINFO. Version 2.0 is now ready, which not only includes licensing but also has an extension which enables a user to extract file data for straight import into an EXCEL spreadsheet. This is performed using the -X command line parameter. Much fun was had doing this including seeing just how big a line could be in an EDIT file. It is big, however the limit is reached when the file has to be FTP'd into a PC. Small changes led to success, without the loss of any information. Also a pricing structure was determined, in which activity our good friends at HP gave good and cogent advice. Thank you. Well now it's off to ETUG in Brighton to see you all, find out what's happening in the Tandem aka HP NonStop world and take yet more of those certification exams. I've yet got to find if when you take he last of them then men in white coats take you away! From a programmers perspective they probably do, however if you are a system's manager you probably get a big raise! Anyway, it is interesting going into previously unexplored corners of the system and discovering facts which you previously did not know and which if you did use you probably would get fired for doing. Like beating up the 2Gb disk drives on the old Himalaya system to get them working, owing to a manufacturing fault which causes them to jam. Something which I cannot find in any manual, thanks Gary, who also told me that once upon time a system manager was just being frog marched from a site and was only saved by the timely intervention of Mr Hall on the line. The correct procedure was to buy another disk from you know who and wait for that to fail as well. Till tomorrow! August 28th, 2007 Product pricing strategies are always difficult to determine. If you cast the price too high nobody will buy the product and if too low the increased sales will not offset the loss of earnings per unit. The Tandem market (sorry HP NonStop) is made considerable more difficult by the fact that there are so many different systems and CPU configurations, DMR/TMR etc. Thus, with ETUG fast approaching and the end of the year, when the free license on FINFO runs out, we need a fairly clear pricing strategy for this product, i.e. base price and multiplication factors for all system types and configurations. Dave Russell and Alan Layzell of HP were very helpful, but the pricing strategy of a hardware vendor, who is trying to sell the latest and best systems, i.e. Itanium, is not probably the best one for a small software house. However, it does give a relative feel for how the pricing should sit and was in fact very helpful. So, thank you very much. We will have the finalised price list available within the next two weeks, together with order forms, so that you can place your orders for FINFO Version 2.0, make payment and get delivery before the year runs out. Thus, work on the licensing program continued, with all the rough edges being ironed out and further research into HP NonStop system identification leading to the discovery of further quirks, all of which have to allowed for. Expect the unexpected! However, that is the nature of a solid product. Finally, last week I got proof positive that good tutoring really does work. The feeling of pleasure when you know that you have helped someone achieve the impossible and thereby their dreams, is difficult to describe. It will be hard to beat that this week. Hoping that you too have a similar week, this week. August 19th, 2007 Last week was a busy one and also a sad one in that I said a final goodbye to our good neighbour Jonathan Wainwright who died at sea on 22 July. Jon had a long history of heart disease which did not stop him being the life and soul of the Old Gaffers Association who are interested in sailing and maintaining gaff rigged craft. Our hearts go out to Margaret his wife and Elizabeth and Andrew their children. Thus I was less than careful about the latest update of my web site and failed to check the link to R-IPPS in the index page. I apologise to all of you for this broken link and will try to be more careful in the future, it is of course now corrected. The rest of the week was spent reviewing the RSI LICENSE manual. It is remarkable how many ideas for product improvement come from such reviews. So I had a great deal of fun trying out the options and putting the changes in to the next release. However as most of these changes will just improve the quality of product maintenance, they are not that exciting. By, the way the flood of interest in backup over IP did not occur, back to the drawing board! Anyway, next week I will be visiting clients and continuing with the process of product re-view. Hope you have a successful week too. August 7th, 2007 I have just finished preparing the copy for our R-IPPS IP Prototyping and Test Suite to be forwarded to Tandem World for publishing tomorrow and making the appropriate changes in our web pages. So I am a little late in preparing my weekly comments about the things which interest me a provider of Tandem utility software. However, during the course of my discussions with providers of Tandem hardware and software solutions, I came across an interesting requirement. Is there anyone out there interested in a secure backup over IP solution? If so let me know! If there are enough of you out there willing to put your pence in the hat then maybe I could write a solution for all of you which will not cost you an arm and a leg, so as to speak. BUT I need to know. Looking forward to the flood of mail, maybe? I'll let you know what happens. Have a good week. July 30th, 2007 Last week I was tidying up and since I was thinking about this web site I entered Google's search page and typed in "Rupert Stanley Ross Systems International" and was surprised to see among the expected pages a reference to an article which I had sent this Spring to Computer Weekly concerning the I35 Legislation. On that same day I opened my paper and read about Geoff and Diana Jones of Arctic Systems and their battle with the Taxman to recognise the enormous amount of effort put in by the so called non-revenue earning partners in small trading companies. The Government really seems intent to kick our industry in the guts. I think that there are two reasons behind this: Firstly, envy, small traders with their enterprise are free to explore commercial opportunities as they see fit and can do it quite well, as shown by the response to the year 2K problem, in a way which the regimented bureaucrats of Whitehall can only dream about. Secondly, control, Information Systems are the key to the Government's ever tighter control of the population, as the rest of us sleepwalk into their Orwellian Big Brother world and free thinking entrepreneurs are not welcome here, unless they are big business which can be assimilated into the States Apparat. None of this is going to help in the long term. The fact is however that there will shortly be no UK software industry, it will have all gone to India or wherever, and consequently there will be no revenues coming from it. The Government should be well aware of the Rothschild's aims as they hung their red shield above their counting house in Frankfurt 250 years ago. Pursue your lust for money or whatever with all you heart, soul and mind and keep it in the family. They are now one of the most successful Banks in the world and the family is still in charge. We invented computers in the UK, we now no longer manufacture them and soon will no longer own any companies able to program them. I'm depressed. I've still got some code to fix. Thank God! Oh! and by the way I also had an interesting conversation with Mike Bond on Friday. A kindred spirit, we both love HSMs or whatever you call them. Till next week. July 21st, 2007 Firstly, an apology. I will not be giving a presentation in October at the ETUG in Brighton. Payments security software using HSMs is apparently not of great interest to the membership of BITUG and ITUG and the last thing they would like to hear about is a case study of such a development project: approach, planning, testing, what went wrong, what could be done better what tools are available to help, how does the future look and so on and so forth. Strange, since so many financial institutions use Tandem/HP NonStop for their payments and other Internet based solutions. Maybe, this is symptomatic of the current state of thinking in these institutions. Anyway, if you are interested, I will still be producing my presentation notes in October for you to download and I will be in Brighton if you would like to meet me. Secondly, the blade server concept for HP NonStop Computers has been a fruitful area for speculation for two good reasons. 1. A
long time ago there was a great deal of speculation about if Tandem would ever
make the NonStop technology work on the Intel Platform, especially the x86
processors. 2. Switchable stacks of blades are a brilliant idea. One moment you are running NonStop on a processor and the next Linux or Windows or whatever. The only thing which is required is a fairy godmother to provide the magical endian switching for the memory management, interrupt handling and dispatching in this crazy mixed up environment. Get real, HP, the only way to do this is by rebooting the OS at change round time, something which I have been doing on my PCs for a long time. Oh dear! Not such a novel idea. Till next week, hope all goes well. July 13th, 2007 Wednesday was a BITUG SIG on S to Integrity NonStop Migration and Data Centre Management. It seems that not so very long ago Tandem were trying to pull their machines out of the server room and into the medium sized businesses of this world. This meant that the machine sizes and power consumption of their offerings were reduced to the scale where they could be installed into an average office, with very modest air conditioning and using standard power supplies. That was the Himalaya K Series. The march away from this objective has now been going on for a long time and does not seem to be losing pace in any way. The prime driver for this is the enormous information processing power needed by modern servers, which finds its ultimate expression in blade servers, set into standard racks, stacked one on top of each other in 6' to 6'6" high units consuming 35kW and requiring air conditioning units of a comparable power to keep them cool. The obvious place to keep this kit, what with the globalisation of IT services, is Greenland or the Antarctic, where at least the contribution to global warming of the cooling systems will be minimal, until the ice melts. The green credentials of this lot are not worth mentioning, but in fact very small compared the the hundreds of millions of PCs across the planet all making their own contribution to the huge IT carbon footprint. The only solution is for the chip manufacturers of this world to make their chips more efficient, which of course they are doing. However, any gain is more than offset but the demand for more and more PCs and the huge server centres to process the tidal wave of data being generated in this wired information age. Anyway, it was a good SIG with plenty of help, hints and practical advice for those wishing to migrate from HP NonStop S-Series to NS-Integrity, which in fact is not a blade server, yet, and surprise, surprise, the needs of the data centres to house them. The next London SIG will be OSS and MQ. The march to the server world of the future involves greater and greater standardisation both in terms of operating systems (Linux/Posix) and the means of communication between servers (MQ). Proprietary systems are slowly dieing, at least until you look under the covers, where you will still find good old Guardian. See you at the next SIG in September. July 4th, 2007 Well, it was not development was it? Running a business means, well, you have to do business things, like talking to people, exploring prospects, getting the accounts in order for the second quarter and a thousand other things which provide the infrastructure for a modern software house. In fact the only software type activity I was engaged in was performing the monthly backup, making sure all the sources were archived in PVCS and shredding old listings. Life goes on. Hope you enjoy your 4th July. June 15th, 2007 I really do not know who learns more the teacher or the taught.
Next week, thoughtfully, back to development. June 9th, 2007 Another quiet week except for the fact that I prepared some copy for Tandem World. Next week I am teaching, so not much doing. June 1st, 2007 This has been a quiet week on the development side since I have been teaching. However, I have managed to update the RSI LICENSE page, as I said I would last week. Next week I teach again, so it will be all quiet. Let me know if you like the new page. May 25th, 2007 At the end of a long period of interesting developments comes consolidation and documentation, documentation, documentation! So, you will see some changes in the web site next week, especially for licensing. Hope you like it, when it gets updated. Let me know your comments. May 18th, 2007 This week has been interesting. The conversion of the Licensing suite into a C++ library is still has been done and tested. It as expected was simply a matter of encapsulation, but to facilitate the ease in which a licensor can add security to their programs a Check public class function was added together with a private re-check class function. This enables a running C++ program to keep on checking the license at certain vital points. This was added because some C++ Programs and kept running for months and years and of course their licenses need checking from time to time. The ease of this interface led to a very simple CheckLicense routine being added to the C library, which does the whole job of checking the license, together with a routine to print license infringements to the home terminal and EMS log, and to point score the number of infringements, too many and you are out! BUT, if you purchase a license from us, you will never be slung out without sufficient warning. The follow on from this was the realisation that the C++ Library
only needed to write infringements to the EMS log after initial start-up. Which
leaves the cleaning up and testing of the licensor suite to be done so that it
is: A side shoot from these activities was to look at how our software suites are divided into groups. So, now since we have so many IP Utilities, these are going to be classed as a separate product to consist of the universal IP Listener and Client and the UDP Terminal. Also, it became apparent that there are a number of utilities which are globally used and to enable checking of these, the wild product check was introduced, so that these universal products like VIEWLICE (the license viewer) and CONSOLE (the universal operator interface) can be used with any of our other licensed products, but the user must have at least one valid license for them to operate. These considerations, together with other business leads, kept me active this week. Yes, interesting, till next week! May 11th, 2007 This week has been a success, the Licensing suite has been modified to allow the licenser to decide if the license is invalid if it is out of date, the CPUs have been upgraded or increased in number, and the license file has been altered to allow it contain the licenses for many different systems. A universal license infringement reporting routine has been added, to send warnings to the output file and EMS log, if an infringement has occurred, and to add up infringement points and decide if the program is to be halted if a user defined maximum has been reached. For instance half a year is one point, each CPU extra is one point and CPU upgrade is two points. This allows a User plenty of rope before the plug is pulled on a licensed program. It is of course possible for a licenser to disregard the infringements if he so decides. The conversion into a C++ library is still to be done, but that is simply a matter of encapsulation of a selection of the C library routines, which is a matter of a day's work. As for the IPCLIENT User Documentation, that will have to be done next week after the C++ Licensing is stable, since it is needed for the licensing of the IPCLIENT package. Finally, if you are going to the BITUG S-Series to Itanium Education day, see you there. May 4th, 2007 This week has been a success, the IPCLIENT program is finished, but the user documentation is still to be written, so no release yet. Adding the operator controls and reporting have made a great difference to the feel of the program. It is now possible to add and delete the $RECEIVE extensions to IP Routing and to mark routes up and down and of course to see what is happening. This all means that once IPCLIENT is running there is no need to pull it down for reconfiguration and also if there is a problem with one of the referenced servers it is possible to get IPCLIENT to automatically re-route the traffic to an alternative destination. All this is needed if IPCLIENT is to be used to route production traffic from applications using SDLC HSMs and Host Interface Processes, to IP HSMs. Dynamic routing, automatic load balancing and decent reporting are highly desirable for a front end process doing this and IPCLIENT does it. So it is back to the licensing suite today and maybe for the start of next week and then on to writing the documentation for IPCLIENT and updating the RSI TEST Documentation, which was always intended to include the IPCLIENT application. Hopefully, by the end of next week I will be able to publish these and put them on trial. The trial licenses will available on demand. I will inform you next week how to proceed, until then have a good week. April 30th, 2007 What a week last week was! It all started so well. Yes, I did complete the testing of the Universal IP Client on schedule and did start on the re-working of the licensing utility and attended the BITUG Security SIG and Infosecurity 2007. However, new ideas tend need time to assimilate and do cause people to re-think their ideas. The BITUG Security SIG on Wednesday gave great pause for thought. The ever increasing sophistication of modes of attack (Phishing, Pharming, Vishing, Man-in-the-Middle and Man-in-the-Browser) and the increasing criminalisation of hacking community show that as we all knew the Cyber Rats in the Sewers are getting more sophisticated and doing it quickly, and showing how out of date two factor authentication is becoming and that the only way forward is shared secret dual path transaction verification and authorisation. The continental banks might have a chance, because they already have very good security in place and have been using smart cards and other tokens for years, but I fear that British Banks with their ponderous and backward looking approach will probably not be agile enough to avoid the security nemesis which is bearing down on them. We will have to see. This briefing , of course made me able to look very well informed at Infosecurity 2007, where to my pleasure I noticed that the UK Security Industry was very well aware of the up and coming problems posed by the these threats, especially man-in-the-browser and are actively coming up with counter measures to these attacks, in particular Validsoft and Tricipher. I also noticed that because of the increasing transaction rates more and more hardware is being thrown at security to replace the previous software solutions. The penetration presentation at BITUG by Colin Yates of TCM and Mike Jones of BrightStrand also inspired me to buy a book and some NESSUS software to find out how to use at least this tool. Anyway, if that was not enough, with a shortened week and only Friday to go CAIL sent me a copy of their Suite 8.0, and what a product it is! One of the actions that I have to perform from time to time is to log onto my system over the net and although the authentication process is good through both the firewall and SAFEGUARD. The session is performed unencrypted, which was worrying me. I would have loved to set up a VPN but that meant having an extra IP Address, which I do not. So the arrival of CAIL Suite 8.0, with its ability to encrypt the session in so many ways was a godsend. Conversations at the BITUG SIG also lead to my realising how useful the Universal IP Client would be in communicating with HSMs. However, the product, as originally written, was only seen as a test tool. In order to make it a viable product, the operator interface needed to be considerably extended to include dynamic configuration and reporting. Also, in the same conversations came inquiries which could lead to further business. It was thus in Friday that I was pulled many ways at once, with licensing pushed to the background. I did what was required, but only just. What a week! Maybe this one will be better. April 20th, 2007 This week the on-site testing of FINFO completed sucessfully. So I am happy to announce that the new release FINFO Version 1.4 is stable. During this test an interesting attribute of FILEINFO and FUP was exposed, in that they display the file partition size on the disk being reported on and not the complete file size. As a result of this I have included the -A optional parameter into the command line. This stands for Actual or All and it results in the complete file size being output over all partitions, a useful addition if you want to know how big your files actually are, if this parameter is not included FINFO behaves as expected by FUP etc. I have also updated the FINFO documentation to include this fact and convert it into a product manual. The testing of the Universal IP Client. (IPCLIENT) also continued and I hope to have it complete by today. Next week is the Info Security show at Olympia in London and the
BITUG security SIG. So the focus once again will be on security. Hope to see you
there. April 13th, 2007 This week the testing of the Universal IP Client. (IPCLIENT) continued apace, however owing to the short week and other distractions it was not completed. Thus do projects slip. As for FINFO, I have received some very valuable feedback from people who have huge systems and files split over several partitions which have exposed its limitations. I have addressed these problems and the new version of FINFO Version 1.3 should now work with these systems. I can't express enough my feelings of gratitude for you users out there who are not only willing to take on and perform the Beta testing of our products but also give us the feedback we need to improve our products. Thank you all so very much. April 5th, 2007 This week at long last I started working again on the Universal IP Client. (IPCLIENT) IPCLIENT is designed to allow anyone to write TCP/IP and UDP/IP Client programs without having a clue about IP communications. All the user sees is a Server that he opens as $CLIP.#DEST, where $CLIP is the Universal Ip Client and #DEST defines the IP Destination. The program is written so each open provides a separate thread through the process and is associated with a separate Socket, either UDP or TCP. IPCLIENTalso allows for STATIC and DYNAMIC routing. STATIC is where there is a permanent Socket object communicating with the remote IP Listener. This is useful if you want to use a service which will allow only a limited number of Sockets, HSMs come immediately to mind. However, it is possible to have more than one of these links all that needs to be done is to provide multiple STATIC entries for the same IP Name, Address and Port in the #DEST mapping table, each one corresponds to a separate STATIC Socket. DYNAMIC is the more usual, one socket per thread type. So, IPCLIENT is now written and ready for testing after about 7 days effort. It contains about 10 Classes and many thousands of lines of code, most of which have already been tried and tested. Such is the glory of using the TELOS multithreaded Applications Framework. The whole application should be fully tested and working by the end of next week. As for FINFO, I tried to tell you all about the new release and got a flea in my ear from a certain quarter. It seems that you don't like being mailed. Sorry for any trouble caused. In future I will keep you updated with developments by means of this page, through the HP NonStop Division and the TandemWorld Publication, unless of course you have mailed me, in which case I will keep you informed with personal e-Mails. On a final note, I found it very useful to use FINFO to search through a large number of sources when I wanted to check where a certain literal was used. I typed in: FINFO -sc101 -sx"literal" and was presented a neat list of source files in which the literal was occurred, by changing the value of "literal" to a procedure name/ class function name, you can find all the files where whatever it is is used. How neat is that? It sure saves time when searching for interrelationships! March 30th, 2007 FINFO Version 1.2 is now available and the documentation has been expanded and completely reworked. FINFO has had a lot more functionality added and it has been put through a thorough testing schedule. Any problems you noticed should now be gone, please tell me if you find any other problems. So, it's been a very busy week. Maybe next week, with the new financial year, I will be able to get back to working on the Universal IP Client, the roughed out code of which has been gazing reproachfully at me from my desk, asking for some attention. Such is life. March 26th, 2007 A problem with the FINFO Search Command -sx"String" and -s("Pira record") has been bought to my attention by Gary Hall, as a result of this I am immediately releasing FINFO version 1.1. To install, FTP it to your system in binary mode and alter the file code to 100. You will find that the command set has been increased, use the -h command for more details. The final touches to version 1.1 will now be released in version 1.2. March 22nd, 2007 Firstly there are two apologies I must make: Firstly, to the Reg Developer for forgetting to put their URL on my site. We have had an enormous number of hits resulting from Phil Manchester article about PIRA but no contacts as yet probably everyone is waiting for the Linux version to be released later this year. Secondly, to HP. Our FINFO is NOT an HP NonStop Utility, it is an Utility for HP NonStop Systems. HP have had nothing to do with its development and have not sponsored it. It merely fills a gap left by the HP FILEINFO command. Sorry HP, I have changed the documentation to correct this error. Now to the news. If you were a member of BITUG and attended the last SIG, you
would have had on opportunity to get a three month license to try out our RSI
TEST product, which includes PIRA, an offer was also made at the same time for 1
day's free consultancy to implement PIRA on your system. Valid for BITUG members
only, sorry all you ITUG members out there but I have to be realistic on travel
times and expenses. Anyway, the original RSI TEST had a cobbled together version
of the IPCLIENT program, which is an Universal IP Client. Which means you can
write a program using $RECEIVE, start IP client and abracadabra, you have a
TCP/IP enabled client application. there are a number of points to note with the
old application. There has been enormous interest in FINFO with a huge amount of
hits on our web site. Anyway, I hope you are all enjoying FINFO as it now stands. Thank you everyone for your interest. March 15th, 2007 Test and Optimisation of FINFO, see below, now complete. It is now ready for release. If you want a trial copy download the Data Sheet and FINFO Program and follow the instructions on the data sheet. The license is valid until the end of 2007. If you like it tell your colleagues and give them a copy. March 13th, 2007 Tandem programmers and system administrators have command of probably the most sophisticated transaction processing system, however in common with many of these servers the control interface for many years has been the command line prompt which is clunky and primitive. HP has now addressed this problem with the release of the Open Systems Management product (OSM) for the control and management of Tandem and other HP Server Systems. Hence, last Thursday ,by popular demand, BITUG hosted their second education day for this well thought out and designed product. However, not all developers and system administrators have access to the facilities of an OSM and also there are still a large number of K Series Systems and if you want to do any form of sophisticated file search the options are minimal, fileinfo and a modified form of grep. So last week I have been developing a new utility FINFO
which enables searches on: I also included the output of user summary results for subvolume, volume and system, depending on the extent of the search. It contains such useful information as user number and name, number of files, bytes used, and percentage use of file extents. I now use this for all my file searches and wonder why I did not
write it earlier. March 2nd, 2007 A lot has happened since I last updated our site, so here is the news: Firstly, yesterday I visited an old friend of mine, Philip Manchester, who I have known since 1973, when I joined Gainsbro Cornard, in Sudbury. In those days he was a Senior Programmer/ Systems Analyst and I was a newly graduated junior programmer. We had a lot of fun in those days writing programs in COBOL on a Honeywell 200 for the textile industry. He was a suave guy in a black suit and I was learning the trade in blue jeans and sweaters. He taught me the importance of documentation, the skills of commercial programming and how to use a hand activated card punch. Anyway, back to business, as you probably know he is a computer journalist and my company is developing test and security tools. Well, amongst our offerings is our test suite which just received a major upgrade with the inclusion of the PIRA Language. The PIRA language makes RSI TEST a very flexible test tool and enables anyone who uses it to perform all sorts of tests on computer systems from simple program functionality tests to optimisation and regression tests on compete systems. The great thing about PIRA is that it enables the user to perform a very large number of tests, varying the content and the transactions rate, with very little script input. This is achieved by using a combination of programmable user hooks and a very flexible data definition language. This flexibility together with the fact that the user environment is informed of the results of the tests during their execution means that the course of the test can be modified automatically as a result of the test findings. Thus it is perfectly possible to write:
What is achieved is the repeated sending of the message, contained in the [ ] and expecting a reply, indicated by the ";" after the ], varying the transaction time from 2000 milliseconds to 20 milliseconds until a stable state has been reached, sampling over 20 transactions to a 1% tolerance limit. How simple can you get! The trick is that the programming elements enclosed by the <> sharp brackets are passed through to the user hook code and that the result of the transaction including response times are also passed to the user hook. This enables a transaction time optimisation to be done, which in effect enables the automatic determination of the actual transaction time. Phil commented that this was great for the quick and accurate assessment of the efficacy of service oriented architectures. Especially as the code is written in pure ASCII C which enables it to be migrated to any platform with the greatest of ease. At present it is implemented on HP NonStop and Microsoft. Windows. The conversation then turned to security and how banks and other financial institutions could protect themselves from the predations of their staff and suppliers. The potential of internal licensing to stop code theft and how to speed up the roll out of security solutions to help them win the war with the hacking and code cracking community, but more about that later. April 5th, 2006 Welcome to our new site and thank you for your patience, it has been a long time in the making, but we do try to keep our promises. It has been the joint effort of Oliver Stanley and myself together with Richard Woods of DEC fame, who did the critical review for me and is an old friend from Essex University. We have also been immeasurably helped by Costi Crallis of Beepweb who are hosting our site. Anyway many thanks to you all. I would also like to say, thank you to all the BITUG members without whose support and inspiration this site would probably still not be published. I especially would like to mention Dave Barnes, Charles Penney, Sean Bicknell, Darren Coffey and last but not least Dave Russell from HP. Of course, a new site is not without its problems, which we are trying to solve and we would be very interested to hear from you about them, together with any other comments. Finally, thank you for visiting. I hope you have found it interesting. Rupert Stanley |
Contact:
info@rsi-ns.com Tel:
+44(0)1206-392923
Copyright © 2006-2010 Ross Systems International Ltd.
Registered in England No.2407494
|