Wednesday, April 14. 2010So Much Changes So Quickly......John Linton .......in this industry it's difficult to keep tabs on it. I was saddened to read this earlier this morning: http://nighthacks.com/roller/jag/ as Exetel's whole business operations are based on MySQL and since the Oracle takeover there has been a lot of doubt as to whether, possibly the most important open source software on the planet, was going to be chewed up and spat out as some over priced 'Oracle Light'. James Gosling, the developer of Java, leaving Sun (recently bought out by Oracle) is one more sign (he wasn't the first) that the MySQL development team is being 'let go' and that is not a good sign for the future of MySQL. Maybe that will take years to happen, if it happens at all, but I think that a company that spends $US1 billion on buying up a company (MySQL) is going to need to justify why it spent the money and why it can justify spending more money continuing to develop it. Perhaps the justification that it is an anti-Microsoft measure is enough but you would have to doubt that. It may seem to be a insignificant issue for someone who has a whole range of more directly threatening issues to worry about but, one thing being in business teaches you is that it's often problems from 'left field' that cause you the most grief as you are totally unprepared for their impact because, like right now, you become wholly immersed in the key day to day problems you have to deal with. All yesterday we had to have the key technical resources of our company trying to resolve a 'bug' that had 'suddenly developed' in our VoIP services which manifested itself in some calls from some customers (and our own VoIP PBX) being 'one way audio'. This was sporadic (perhaps one call in four either inbound or outbound) in our own operation and some different parameters in our customer's usage. As we had made a very minor hardware change the previous Monday we assumed that we had somehow made an error in that minor upgrade so the first thing we did was to 'roll back' the changes we made - it didn't help. By 4 pm we had found what we thought was the error and changed a key component which appeared to resolve all the customer problems but not our own so we concluded, logically but wrongly as it turned out, that, some how, we had induced an intermittent error condition in to our Asterisk based VoIP PBX. The hours ticked by with several engineers getting increasingly tired and increasingly frustrated as we had made no changes to our own internal VoIP hardware and software but were 'convinced' that the problems must be related to the hardware changes we had made in some way. Like the proverbial 'bleeding obvious' it was yet another example of Sod's Law which was finally traced down using our Asterisk engineering/programming resources in Sri Lanka, Chile and Australia as a defect in one of our voice carrier's services that didn't only affect us but affected all traffic that followed a particular transit path either to or from our and our customers end points. The problem was finally resolved a little after 1 am this morning and the carrier advised of what we believe their problem was. While I received confirmation that the carrier has 'accepted the ticket' and is looking in to the reported problem we can't be certain that was the root cause of the troubles we and some of our customers were experiencing but switching that traffic to another carrier eliminated the issue and was thoroughly tested. So that's a 'left field' type of problem that after 3 years of fault free operation comes 'out of the blue, potentially was affecting almost 20,000 customers and prevents Exetel reliably receiving and making telephone calls for almost 24 hours. While I was greatly relieved to receive the news earlier this morning and feel a lot better than I did yesterday it was a salutary reminder that no matter how well something has gone (in this case for over three years) anything can suddenly happen that changes your operating scenario so completely you are completely shocked....and, as in this case, all the redundancy of alternate carriers, duplicated hardware and standby hot spares made not the slightest difference. So 'death and destruction' in the commercial world can come 'out of the blue' from events as varied as a huge US company taking over your source of all your automation and possibly changing your cost structure 'overnight' to a key service that has performed faultlessly for as long as you have had it in operation suddenly failing for no apparent reason with no-one able to diagnose what has gone wrong - dealing with sudden changes in competitive marketplaces or supplier policies is child's play compared to being confronted by events that appear randomly and over which you can exercise no control and have no ability to provide input/assistance to their resolution. Is today going to be a better day? 10.........
Trackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
Well, today I should hopefully not be afraid of calling people with my MoIP then. But it was very frustrating for me and my calling parties over the last couple of days. I think it started for me on the weekend, from my poor poor memory.
Comment (1)
The VoIP issues wouldn't have affected your MoIP and were only extant for something less than an hour on Monday.
Comments (4)
As per The Good Guide (Hitchhiker's that is), there's still legs in mySQL but what's more, the writing has been on the wall for it in terms of bloat & DB evolution anyway.
I think you've got at least 6-12 months of "just fine" mySQL use, probably more. By then, the evolution from mySQL will be much clearer & options like CouchDB, noSQL methodologies & others will be more viable. Don't Panic on mySQL, despite those close to it running around waving their hands in the air & yelling "The end is nigh!". As my old Army supervisor would say, "Stay on target, stay on target!". Comment (1)
It's is sad, but not surprising, to see James Gosling leaving Sun/Oracle.
His comment regarding the IP of blogs imposed by Oracle show the difference in mentality between the two companies. To me (someone who has worked with/for Sun since the early '90) it is clear that Oracle is going to kill most of Sun's great products and innovations. They have already killed Open Solaris and are now charging for Solaris, which means a certain death of the OS. The previously freely downloadable Java Enterprise System (aka SunOne, iPlanet & Netscape server products) will certainly follow soon. I'm not sure how they can kill an open source product like MySQL, as it can't become closed source as far as I know. However they can disrupt the development/progress by the uncertainty around the product. Time will tell I guess. Comments (2)
It was the best for Scott McNealy, but not for any of the customers (both paying and non-paying).
Not good either for the employees or anyone who provides consultancy around Sun's products. Comments (2)
As Edwin alludes to, the good thing about open source is they can't take it away. MySQL is dual licensed, GPL being one choice. So you can rely on what you have now - unless you have a service contract. What you can't rely on is any further development. However given the momentum of MySQL, it is just as likely that one of the current forks or a new fork will take the lead and it will live on - perhaps under a new name.
Comment (1)
You are correct in that they are unable to "close-source" mysql as it currently exists.
The GPL license means that everything that has already been released, stays released. The only thing they can do is close-source it going forward, meaning you don't get any new features (and potentially security fixes). While this is a very real possibility, I would count on the most recent open-source version of mysql being taken and forked off into another community driven project, and the mysql community at large putting their support behind this version. Oracle are free to charge/license their version of mysql however they like, but I have a feeling there will always be an open source version somewhere... Comment (1)
Hi John,
This is the exact reason why businesses are still refusing to take up VOIP. These sort of technical issues just dont happen on a 60 years old dinosaur PSTN. Comment (1)
....which is why all end user systems always involve fail over back up.
In three and a half years we have never had an operational issue. In the previous three and a half years with Telstra ISDN we had one 36 hour outage when a fibre cut occurred and a 5 hour outage when a thunderstorm fried the building MDF. Comments (4)
|
Calendar
QuicksearchArchivesCategoriesBlog AdministrationExternal PHP Application |