Autonomous database

  • April 3, 2018
  • Ok, this is kind of long. Go get a cup of coffee.

    Amid the anxiety and revelations of the Russia scandal including the Cambridge Analytica story that showed how easy it was to steal 50 million Facebook user profiles, it’s easy to mix up cause and effect. Importantly, Facebook wasn’t hacked or broken into but it was used as it was designed.

    This has led some to question whether Facebook as such can exist at all in our pluralistic society while others believe the problem of surreptitious psychographic profiling will blow over once everyone plays by the same rules. After all, others have argued, other entities do the same thing. They point to Google, Amazon and even the traditional print industry as culprits for gathering personal data for analysis and, it should be said, weaponization.

    Of course, the issue is manipulating and weaponizing the data. If we can’t trust the data, then we are disassembling one of the pillars of democracy, the acceptance of scientific rationalism. Boiled down, it means facts are facts even if you don’t like them.

    If you remember a time before social media when identities were not so readily stolen and you think that reality was good, you might also recoil at the thought that those were the good old days, that things are now permanently different. There is a third option though and there are probably many that seek to balance the benefits of new technology with the protections we’ve grown accustomed to.

    This article can’t be all things to all those people but it attempts to find safe harbor in a storm and therefore makes accommodations. If we can’t live with the compromises, perhaps it can at least point out some of the major obstacles to be over come.

    Business model

    It is an article of faith that Facebook’s business model, as well as those of other social networks and search engines, is selling advertising. But it is my contention that this model has run its course. It was effective when the companies were smaller, when their consumers were more innocent to the ways technology can be used for both good and ill purposes. The advertising model was even necessary in a time when the Internet was new and finding people and things was strange.

    The advertising business model was a default that data aggregators took on the way to phenomenal profits and who could blame them. The tech sector has a habit of minting money and the founders of social media and search engines were merely the latest in a long line of prolific brainiacs who struck gold. It is hard to believe that any human in a similar situation would act much differently.

    The latest dustup that dragged social media into the political spotlight now presents two choices to these businesses. They can hobble their products, which could reduce the amount of data they collect making them less interesting to advertisers, or they can change their business models slightly to prevent unethical use of their networks.

    Disruptive innovation

    Anytime a new technology reaches market, it has the possibility that it will disrupt the existing order of things. Disruptive innovations have coexisted with Capitalism since its origins in the Industrial Revolution. Disruptive innovation means making thread and then cloth with high-speed mechanical means, making a steam engine powerful and small enough to be mobile, or making a computer that could fit on a sliver of silicon about the size of your thumbnail.

    The world changed with each of these disruptive innovations and others, because they immediately made an old order irrelevant and they organized whole economies and even civilizations around new driving forces. The Internet and its children are the latest innovations that have rocked the world. In each, humanity has had to grapple with both the benefits and the deficits of the innovations.

    So far, we’ve benefitted enormously from these innovations but recently we discovered their less sanguine side. If history is a guide then regulation in some form is a likely next step. Some leaders in congress have already broached the idea on several occasions but it’s important to get the idea right before pulling the trigger, which is why we need to discuss business models.

    Regulation?

    Regulation could happen in social media and search but there’s much that the technology companies can do to either avert it or ensure that its mandate is as light and congruent with company interests as possible starting with the prevailing business model.

    Although the advertising business model has served many companies well, they’ve morphed into data companies with big responsibilities for safeguarding the data they collect and that’s not something they’re eager for.

    The big data gathering companies like Facebook, Amazon, and Google and their competitors, have become data companies first and advertising vendors second and if this understanding had been realized sooner, many data breeches would in all likelihood have been thwarted. Rule One of business is never give away your product, it’s what you charge for because it pays the bills. Applying the rule should be as obvious as encrypting user data in this case. Additionally, no expectant user of the data should be able to access it in its unencrypted form without, of course paying, but more importantly presenting valid credentials and stating a beneficial and productive purpose of the use.

    I’ve written before about credentialing and how it’s actually harder to pull permits to remodel your kitchen than it is to advertise any message you want on social media so I won’t perseverate. So let’s turn to encryption.

    Security as a business model

    Social and search’s business model must turn from advertising to data management, curation, and selling access to it and we live at precisely the moment when these activities are possible on a very large level. This includes encryption and the same form of certification that applies to other professionals from doctors to beauticians and plumbers.

    Encryption and its reverse take time and require compute and storage resources which have often cut short discussions involving them because of cost considerations. But new, shall we say disruptive, innovations in computer hardware and software are reigniting the discussion.

    In hardware data storage was long accomplished with the hard drives of most computer systems. Data enters and leaves storage on millisecond time scales, which is very fast. However, computer CPUs and memory operate one million times faster at nanosecond speeds. CPU chips spend a lot of time waiting for data to become available even when, as most modern computer systems do, there is memory caching for frequently used data.

    Innovative hardware designs now offer solid-state memory devices that replace disks. This memory operates at nanosecond intervals and eliminates the lag time of older mechanical systems. What should we do with all of this newfound speed? One possibility might be to dedicate a small portion of it to encryption. Typical encryption modes on the market right now could be broken but that would take so many years that the resulting data, when finally available, would be useless and encryption is getting better.

    Encryption would be a good thing but it wouldn’t solve all problems and securing our information infrastructure so that it operates more at utility grade, requires other changes. Bad software, malware, viruses, Trojan Horses, and the like may still get into systems.

    Mark meet Larry

    As luck would have it free markets generate inventions faster than they can be adopted. Often a disruptive innovation exists at the nexus of several disruptions that just need one more critical piece. That’s the case with many of the system level inventions that Oracle has brought to market over the last several years. They’ve pioneered important developments in solid state storage, encryption, chip sets that weed out intrusive malware, and a self-patching autonomous database that just hit the market.

    All of these things turn out to be essential to safeguarding data which will enable the information revolution to continue burrowing its way into our lives and enriching society. They are also the underpinnings of a new business model that turns big data companies into ethical data providers. They might also continue being social media companies but the data tail would now be wagging the dog.

    My two cents

    What do I know? I just read and write a lot. But what I see is an industry about to be regulated and, in my mind, the smart play is for the social media companies to lead the charge to ensure they arrive at something they can live with instead of remaining aloof and having some regulations imposed on them.

    There’s a wild west mentality in Silicon Valley in which what isn’t proscribed is encouraged. But we should keep in mind that the west only remained wild until the pioneers arrived and established towns with roads, schools, and churches. The wild bunch might have disliked the idea of settlement, they might have opposed it, but they were quickly in the minority and civilization won. That’s what’s happening in tech today and we all need to seize the moment.

     

     

     

     

     

     

    Published: 7 months ago


    I spent the first week of the month in San Francisco at Oracle’s customer event, OpenWorld 2017 and when I wasn’t drinking from an information fire hose, it was alternately fascinating and exhausting. There were major announcements in database, blockchain, AI, cyber security, and other stuff I have only a tangential association with. For instance my eyes glaze over when they start talking about bare metal servers and going serverless so I’ll limit these observations to database and security with a few dashes of other things.

    But primarily OpenWorld 2017 was a coming out party of sorts in which the culmination of billions of dollars of investments in advanced technologies became unavoidably visible. From here on Oracle might still support its legacy customers but make no mistake, they are legacy and the future is at minimum about moving the datacenter to the cloud and moving your traditional licenses there. Oracle even has a name for that practice BYOL or bring your own lice

    Oracle is making that task as easy as it can be and I spoke with customers who were doing it or just finishing who were pleased with the experience. We don’t often think about it if we keep to the big picture but there are a lot of old systems in the enterprise that need to be replaced because they no longer support their missions. There is no better indication than that for determining that the next few years will be full of stories about cloud migration.

    The key to successful and necessary transfer to the cloud is Oracle’s autonomous database, 18c. The new product will be available by end of year and is supposed to self-provision, self-maintain and self-patch while running thus eliminating the need for most downtime. CTO and founder Larry Ellison said that it takes an average of 13-14 months from when a patch is available to when it is installed in a majority of customer shops. If a patch covers a security problem that leaves bad guys a huge amount of time to steal data, which has become epidemic.

    As Ellison said in his opening night keynote, “People are going to get better at stealing data, and we have to get better at protecting it.” Fair point. The database will be a big part of that protection along with some semi-autonomous security software (that will become autonomous soon enough). Both products rely on Oracle’s AI and machine learning tools as well as advanced database hardware.

    That’s the catch, but it seems eminently reasonable: you have to be on Oracle gear to get the full benefits of the software’s power. Actually it’s less of a catch than you might think. While there are enterprises that are big enough to need and to purchase the hardware, many if not most, customers will receive the full benefits of Oracle’s technology as consumers of its cloud services. So many of these announcements can be rightly seen as further inducements to move to the cloud.

    Ellison is fond of saying, “You can get all of this, but you have to be willing to pay less.” That’s fair if you’re looking at the monthly or annual subscription charge but one suspects that over time many companies will be paying more overall though there’s the issue of refresh that many companies avoid but which are a standard part of the cloud. We’ll see, in the end you get what you pay for.

    Oracle also announced a foray into blockchain, the distributed ledger technology that provides greatly enhanced security, speed, and transparency to inter and intra-company transactions. For instance you will soon be able to use blockchain to track the provenance of parts in a supply chain, and one can only hope that credit reporting adopts similar safeguards in the future. We can also hope to track some customers and their purchases that way, especially in a B2B setting to facilitate sales.

    Lastly, there’s AI and data. Whether it’s called AI or machine learning, the technology requires lots of data to train a model to be useful in predicting the future. Most enterprise data is deficient in one or more axes of data on hand so Oracle’s solution has been to provide clean data to augment private data and deliver the big picture view that’s needed in sales, marketing, service and a lot more. Oracle also introduced a new set of IoT applications aimed at specific business outcomes. It’s impossible not to say more about this later.

    But to summarize, Oracle’s long-term investment in cloud technology has begun to pay off. We’ve seen this in the company’s earnings reports over the last year and OpenWorld was a kind of coming out party for numerous solutions at literally every level of the software (and hardware) stack. The company will be a formidable competitor in the years ahead as its legacy base is up for grabs. They’re all going to move to the cloud at some point and Oracle wants to keep them. Other vendors with good solutions are competing at every level from Amazon and Google to Microsoft, SAP, and Salesforce. At Oracle OpenWorld 2017, Oracle more than made its case. Time for the others to step up.

     

     

     

    Published: 12 months ago