Pervasive

February 2010 - Posts

• Cloud Data Challenges Require New Cloud Data Answers

  Congratulations Twitter team!  50 Million Tweets!  That’s a lot of data.  You would think it would be an issue for the folks at Twitter to take care of all that data. Actually they seem to have moved past their issues. I for one have not seen the "fail whale" in some time.  At the very least I have not seen him as frequently.  There was a time when they struggled to keep their heads above water and the whale in the water.  Remember SXSW 2007?  All this data highlights a point, data on the cloud is growing, and FAST! And it is different than anything we experienced in the past. For people who work with that data it means we cannot approach it the manner we approached it last year, or even last month.  We need to think in new terms, relative to what is actually happening on the cloud.  Huge data volumes, but not all the time. Look at these charts.  The first one is the volume of Tweets during the Super Bowl against time. The next one is the volume of Tweets against time for the last few years. Two deductions can be drawn, the volume is huge and growing, but it is not linear during the day or events.  In the old days when asked to process that data, we would have provisioned just enough static servers to handle the peak load and written off the “at rest time” as the cost of doing business.  We can’t afford to do that today.  Cloud infrastructure is available and should be leveraged.  I have seen vendor after vendor throw their software “stack” from last year on an image in the cloud and call it scalable.  Really?  Ask them to explain their "scalable infrastructure". Ask them if you had 1TB of data arrive over a 2 hour time span could they scale and process it without manual intervention?  If they say yes, ask them to prove it... If you have varying volumes of data that are growing, and/or a varying demand to process that data in real time, on-demand or in even in batches on the cloud, then ask those questions and make that vendor stand behind it.  …it will be too late if you find out during data crunch time.   Posted Feb 23 2010, 01:40 PM by bjacaruso with no comments

• The “as a Service” Argument in Cloud Computing

  I have been called argumentative at times although I like to think of it as being controversial.  I will also say I like a good discussion as much as the next deliberator, however, there is a debate raging in the cloud space that I just don’t get.  The basic question in the battle field is what vendor service(s)is classified  as either Infrastructure as a Service (IaaS), Platform as a Service(PaaS) or Software as a Service (SaaS).  In our world we look at it like this; The Pervasive DataCloud2, an example of PaaS, it sits on Amazon Web Services, what I consider to be IaaS, and provides services to IaaS, PaaS, as well as SaaS. I could be wrong, but come on folks, does it really matter what label is on the vendor?  If they have a service that fits your need, will you not buy it because they (or more likely some analyst) labeled it incorrectly?  Check out the latest thread on the Google Cloud Computing Group . This argument is gaining in size and seemingly in importance to rival what “size of government”, “amount of taxes” and “the national debt” already own in the political circles of this country. I am pretty sure if my electric company changed their classification from an utility to a Electricity as a Service (EaaS) I would not stop using the current that comes out the end of their wires.  I am also sure that someone will take offense to my simplification of this description, but if you do, please help me by explaining the importance while you’re complaining.     Posted Feb 21 2010, 08:06 PM by bjacaruso with no comments

• Software to Cloud Service

  Adobe announced their migration of their Life Cycle application to the AWS cloud. They didn’t just copy the old application, they migrated the idea into a new service.  The key word is service, software as a service, but a service none the less.  (I wonder if we should start calling electricity utility as a service UaaS). Independent Software Vendors (ISV’s) are realizing that simply cutting and pasting their old apps onto instances of VM’s in the cloud does not make a company a cloud software company.  More thought and consideration must be put into the application, and while pieces of the intellectual property can be used, some work is required to leverage the goodness of the cloud.. There is a distinct difference between a cloud service and software that can be leveraged on virtual instances.  Be very careful of the latter, it probably does not scale or take advantage of the things you moved to the cloud for in the first place.  True cloud services should and do take advantage of the cloud.  They expand and contract like the cloud.  They are available for use anywhere, and they can be leveraged from any cloud. Be sure to ask is an ISV's software meets these requirements BEFORE you sign up for a cloud service that is not a service at all, but simply repackaged client server code.   Posted Feb 10 2010, 03:17 PM by bjacaruso with no comments

• Computing Cloud Price War Breaks Out!

  This is one war we can all be happy to see start.  Microsoft and Amazon have fired some warning shots in the past few months, but yesterday Amazon sent a clear message to Redmond that this is a war and they are ready to fight. That’s awesome for us the consumers.  Regardless of where they finally drive the price, we can have some sense of confidence that the money we pay for cloud computing cycles and storage are now being driven by the forces of a fair market. Adam Smith would be proud today, because the market is at work balancing itself without the intervention of outside forces. Only time will tell how this plays out, but my guess is that prices for cloud computing will move slightly lower over the course of the next 6 months.  At some point it will stabilize around 10 to 20% lower than we see today in certain cases.  It is unlikely that we will see prices take on the Moore’s Law effect some have associated with the cloud phenomenon. The short message is, if you were worried you were going to get price-jacked at the entry ramp to the cloud, you can set that fear aside. On a closing note I have a prediction that we will see someone attempt to produce a clearing house for cloud purchasing.  Hiding the complexity of finding the lowest price, farming simple jobs to cloud based computing for that day, week or month. It will resemble a stock exchange or hedge fund implementation for cloud computing resources.  Think of it, a company built around the idea of leveraging future prices based on market events and predictions.  Oh yea, they tried that with energy didn’t they, it was called Enron…   Posted Feb 04 2010, 09:01 AM by bjacaruso with no comments

• Don't bring your junk (dirty data) to the Computing Cloud

  I decided to clean out our garage over the weekend.  As I sorted through the now outdated lawn sprinklers (we have a sprinkler system at our current address) and light bulbs that worked in fixtures on previous porches, it occurred to me we moved a bunch of junk into my new house.  The news is we have lived at this address for 8 years. I sure do wish I had taken the opportunity to go through this stuff when we moved.  What this means is I have been picking it up, moving it out of the way, to find what I was really seeking, for eight years!  What a waste of time. The scenario above repeats itself all over the world every day.  But it is not limited to urban dwellers and garages.  As organizations move their data into the cloud, the potential for the same “garage full of junk” exists.  Dirty data files that organizations have struggled with for years are now being transferred, in bulk, to the cloud.  And just like our garage full of junk, those organizations will continue to stumble over them for years to come. I would suggest that organizations transferring data into the cloud, plan an effort to identify and clean their dirty data before the move.  Using profiling tools and customer data services that are positioned “inline” with the flow of data coming from your on-premise data stores to the cloud is not just a good idea, it is a strategic advantage over anyone you compete with. Regardless of how you move the data (bulk file transfer or migration and transformation services running on  Pervasive DataCloud2) take the time to go through your junk! Leverage a data service to profile and then even remediate your data.  Otherwise plan to spend your Saturday like my last one, sifting through junk!   Posted Feb 03 2010, 09:11 AM by bjacaruso with no comments

• Action Items for Security in the Cloud

  Cloud computing is growing and will continue to grow for the foreseeable future, just Google Cloud Computing Growth to confirm.  But is it growing as fast as it can?  Probably not.  Security continues to be the number one concern, simply Google Cloud Computing Security to confirm.  There are many issues in any discussion about cloud computing and security, but this article in ComputerWorld caught my eye this morning:  http://bit.ly/SecurCompWorld. Elisabeth Horwitt captures some right now action items that cloud consumers can exercise to help insure their cloud effort is more secure. Posted Feb 01 2010, 08:26 AM by bjacaruso with no comments