Loading ...
Sorry, an error occurred while loading the content.

Re: [decentralization] bigcorp winners and losers

Expand Messages
  • Rahul Dave
    Ok, I m going to stick my hand out and take a hand at making some predictions and observations... ... Did Microsoft really losse in the web period? My claim is
    Message 1 of 9 , Aug 25, 2000
    • 0 Attachment
      Ok, I'm going to stick my hand out and
      take a hand at making some predictions and observations...
      I got this from you:
      >
      > Let's say there is a post-web future where users are generally asking the
      > internet for functional specs ("give me this...") rather than looking for a
      > provider and then looking for a resource at the provider's site, and that
      > this leads away from the current focus on specific sites. Let's say that
      > there is a parallel trend to do even more work on the internet, rather than
      > your local machine, than you do now. What does this mean for bigcorps?
      > Some examples:
      >
      > * Sun and Oracle have done well at Microsoft's expense in the web period
      > because web servers have taken over work from PCs, and Sun and Oracle are in
      > their element on web server-scale boxes. But huge collections of smaller
      > boxes collaborating Napster style avoids the need for huge single boxes.
      > Would Sun, Oracle et al lose market share? Would Microsoft take it back?
      Did Microsoft really losse in the web period? My claim is they didnt, and
      it is precisely them and Intel which enabled the explosive growth of the web.

      The point is, left to anybody else, IBM, Sun, etc, the prices for PC's would
      not have been so low thus enabling a lots of people to get onto the web.
      The downward spiral of prices has a few reasons: the relentless MS Office
      driven upgrade cycle, oversupply of components, MS OS's being priced
      reasonably, as opposed to the proprietary unixes.

      At the same time linux started developing, and development took off majorly
      at about the same time the internet did. So linux owes one to microsoft,
      strangely enogh for having built such a base.
      And the growth of the internet owes a lot to linux in terms of modem
      boxes for ISP's and 40 percent or so of all web servers. This trend
      will probably continue into the appliance stage(kerbango radio, etc).

      And Sun, got to sell more servers and oracle more databases.

      As DSL/Cable takes root. I see the PC becoming eminent again. Everybody
      has a server. The server will serve both the house as well as the outside
      world. It may have a firewall for filtering porn/viruses/whatever depending
      on the family's choices. It will allow configuring home appliances.

      I wonder if there would be a market for thin clients of this server in the house? I think not, but I certainly think this server could be helpful in
      backup situations.

      I think each such PC/server will have a backup/cache peer at the ISP, so that
      the user can access from outside even if the home network is down. For
      56K users, this will be the main server, not the home machine.

      Synchronization will be part of the system. What you wite on your laptop/PDA
      will be synced to this "server" machine. You will be able to
      put some content/services here for public viewership, and the rest will
      be secured for access to your family group. The group access will be
      PKI certificate based, instead of usernames passwords.

      Groups of users will be able to create their own p-p subnets, to restrict
      to their families or fellow professionals. One server will take part in
      multiple subnets, and there will be some file crossover.

      Households will not be the only people on this system. Websites will
      morph into nodes on the system. Some websites will only be accessible to
      specific subnets.
      Subnets may have their own ads, knowledge aggregation
      mechanisms, multicast policies..searches might be more direct.

      These will need to be powerful nodes as many requests will come to these ones.
      With agressive caching in the subnets, not all parts of the service will
      need to be local, so the database needs may be mitigated some, but they
      will still be there. So the present demand for big-iron will level out
      but always be there.

      One interesting possibility maight be the dilution of brand. Presently
      with the volatile stock market, only the strongest brands survive. These
      are not cecessarily the best, but they are the ones who spend most on
      marketing. If epinions style ratings became part of the browser, along
      with google style link/use statistics, service fulfillment at smaller
      houses may get higher on the reputation chain.

      Online adverisement might change. Presently you get stupid ads at websites,
      but the owner of a "server" in this new system may publish what kinds
      of ads he/she wants, eg: computers, atleast 10% discounts. These will be
      the ads that he gets shown. All this is possible since the browser is a server
      and thus queryable.

      Microsoft is well poised via .Net for the services part. They havent got
      p2p yet. But they are setup for it. Win2k ships with Microsoft Message
      Queue(MSMQ) and IIS Personal web server. .Net has ASP+ which can run inside
      IIS as well as IE. ADO+ allows for data caching into XML datasets which
      can then be moved around nodes..(for privacy we can use a symmetric SSL
      session key).

      MS will need to bundle SQL server into the OS I think, and open source
      relational db's like mysql and postgres, and object ones like Zope, Frontier
      will
      find themselves on desktops, and smaller versions in handhelds...
      >
      > * Given the emergence of p2p auction services, can eBay hold its value?
      >
      EBay will probably morph into the auction service cache! Its the reverse
      bandwidth issue. They will be a node on a P2P network getting
      the specified auction content, and because of their big pipes be a preferred
      node(sooner or later p2p systems will need to do bandwidth ratings)
      > * Napster is obviously messing up the AOL/Time Warner merger. Let's say
      > this is a reasonable problem, where a new media company stands to win market
      > share by cannibalizing a business (music distribution) that has historically
      > higher profit margins. Does this mean that other bricks & clicks mergers
      > are doomed, because of the cannibalization problem?
      I think this will depend on a case to case basis. However, it seems to
      be that high profit margins may themselves be doomed. This may not
      be a good thing in itself, but a positive outcome is that there may be
      a greater alignment between quality(and reputation) and profit.
      >
      > * are there other bigcorp trends in the post-web internet?
      >
      See above for ramblings :-)

      A lots going to depend on how browsing changes, and who does
      both p2p and data synchronization. The killer app today, as Dave
      has said before is music. But for the large p2p to succeed, its going to
      need support in the browser. But the big browsers: AOL/Netscape and
      Microsoft/MSN are stuck in the portal-gatekeeper-proxy to internet mentality.

      Will Microsoft have the courage to break free? Can we turn mozilla from
      just another browser to a killer set of technologies(XPCOM, gecko,
      access to DOM) for the p2p internet? Will there be as usual two
      competing situations? MS:SQLServer+MSMQ+IIS+IE+Universal Canvas
      and OpenSource:MySQL+?+Apache+Mozilla+GnomeCanvas?

      All intersting questions. My claim is, user interface is key. P2P should not
      be harder than browsing, and its one of the reasons Napster is so good I think.

      Rahul
      > - Lucas
      >
      >
      >
      > To unsubscribe from this group, send an email to:
      > decentralization-unsubscribe@egroups.com
      >
      >
      >
    • Lucas Gonze
      Let s say there is a post-web future where users are generally asking the internet for functional specs ( give me this... ) rather than looking for a provider
      Message 2 of 9 , Aug 25, 2000
      • 0 Attachment
        Let's say there is a post-web future where users are generally asking the
        internet for functional specs ("give me this...") rather than looking for a
        provider and then looking for a resource at the provider's site, and that
        this leads away from the current focus on specific sites. Let's say that
        there is a parallel trend to do even more work on the internet, rather than
        your local machine, than you do now. What does this mean for bigcorps?
        Some examples:

        * Sun and Oracle have done well at Microsoft's expense in the web period
        because web servers have taken over work from PCs, and Sun and Oracle are in
        their element on web server-scale boxes. But huge collections of smaller
        boxes collaborating Napster style avoids the need for huge single boxes.
        Would Sun, Oracle et al lose market share? Would Microsoft take it back?

        * Given the emergence of p2p auction services, can eBay hold its value?

        * Napster is obviously messing up the AOL/Time Warner merger. Let's say
        this is a reasonable problem, where a new media company stands to win market
        share by cannibalizing a business (music distribution) that has historically
        higher profit margins. Does this mean that other bricks & clicks mergers
        are doomed, because of the cannibalization problem?

        * are there other bigcorp trends in the post-web internet?

        - Lucas
      • mark
        ... One of your basic assumtions is flawed. It is that the growth in the availability of bandwidth with outpace growth in storage capacity. Storage capacity is
        Message 3 of 9 , Aug 26, 2000
        • 0 Attachment
          > Let's say there is a post-web future where users are generally asking the
          > internet for functional specs ("give me this...") rather than looking for a
          > provider and then looking for a resource at the provider's site, and that
          > this leads away from the current focus on specific sites. Let's say that
          > there is a parallel trend to do even more work on the internet, rather than
          > your local machine, than you do now. What does this mean for bigcorps?
          > Some examples:

          One of your basic assumtions is flawed. It is that the growth in the availability of
          bandwidth with outpace growth in storage capacity. Storage capacity is continuing to
          grow at such a rate while capacity prices continue to drop that the likely future
          will be one where the average home will have much more than a terabyte of
          information storage conected to a variety of devices, and the trend will continue
          to the point that it will probably be faster and cheaper to load a device containing
          a copy of all the information currently on the internet into the back of a truck and
          just drive it to where you want it. It doesn't take a significant advance in
          technology to reach this point. HD-ROM is just around the corner, thats 160GB on a
          CD size disk.

          > * Napster is obviously messing up the AOL/Time Warner merger. Let's say
          > this is a reasonable problem, where a new media company stands to win market
          > share by cannibalizing a business (music distribution) that has historically
          > higher profit margins. Does this mean that other bricks & clicks mergers
          > are doomed, because of the cannibalization problem?

          I don't see any indication that Napster or P2P technology is or could interfere with
          AOL/TimeWarner and the business model. Napster type technology may be the earliest
          indicator that information distribution will not be a viable business for much
          longer. The problem is that AOL/TimeWarner is not exclusively or even primarily a
          distribution business. They are in the business of producing material on behalf of
          sponsors(advertisers), and they are in the business of providing a distribution
          medium on behalf of the distribution businesses(ie: cable networks).
          AOL and Cable networks and other business in the distribution will probably suffer
          as the availibility of bandwidth and storage capacity increases, but the core
          businesses will not suffer.

          It is necessary to face the fact that only a very large centralized production
          company can produce entertainment of the quality and quantity we now expect. Perhaps
          in some distant future, we will manipulate digital actors on our PCs to produce
          movies or television superior to that which we have today. But this is likely to
          prove a market that only a very large centralized production company could take
          advantage of. Big business wins either way.

          We also have a great deal to loose if decentralized media becomes the de facto
          standard. Decentralized news is just rumor mongering.
        • Rahul Dave
          ... I think that the web has brought the reverse effect to the fore already. Atleast in the written word or news, I can get content tailored to me, unproduced
          Message 4 of 9 , Aug 28, 2000
          • 0 Attachment
            > > AOL and Cable networks and other business in the distribution will
            > > probably suffer
            > > as the availibility of bandwidth and storage capacity increases, but the core
            > > businesses will not suffer.
            > >
            > > It is necessary to face the fact that only a very large centralized
            > > production
            > > company can produce entertainment of the quality and quantity we now
            > > expect.
            I think that the web has brought the reverse effect to the fore already.
            Atleast in the written word or news, I can get content tailored to me,
            unproduced by CNN or the like. The deal is, I can aggregate information
            from various small sources who I find trustworthy, as the barrier to
            producing content is lowered.

            In the old days I had to take what CNN gave me, or NY times gave me, with
            all its biases and assumptions. While the big news corp still wins today,
            because its breadth is unequalled, its target demographic is not me, and
            I am not orphaned today as I would have been in the past. I can go elsewhere.

            So there is the tragedy of commons that capitalism brings due to the necessity
            of targeting to your commom market. With the web, we can aggregate to route
            around it. Where P2P is important to me is the (a) groupware, common bookmarking
            aspect where I can find my demographic and run with similar people, and
            (b) it provides a visibility channel that the web failed to provide: a small
            news shop can now compete with a big shop on a reputation basis to sell
            the same news, or product, and have a dedicated following from over the world
            and be viable.

            The question then, for me, is, can we engineer a P2P system which allows for
            a distribution channel which is not controlled by advertising or marketing,
            but rather by a reputation system, with reputation evaluated in sub communities
            and then the community at whole, both by active rating, link aggregation, and
            answering the question: where do reputable members of this community and
            the people on the whole surf. All without violating privacy...
            Rahul
            >
            > I don't completely agree - weblogs are both compelling, popular and
            > decentralized - but I see your point. The big content makers dominate for good
            > reasons. On the other hand, the good reasons are mainly about economies of
            > scale driving competing creators out of the distribution channel.
            >
            > - Lucas
            >
            > (p.s. sorry about semi-random line breaks above. ...have recently been forced
            > to switch to Outlook, which does hell to replies).
            >
            >
            >
            >
            > To unsubscribe from this group, send an email to:
            > decentralization-unsubscribe@egroups.com
            >
            >
            >
          • Lucas Gonze
            ... Hmm. I m not sure that what we are farming out to the internet is storage. Rather than dumb hard drive capacity, it is decisions and expertise. For
            Message 5 of 9 , Aug 28, 2000
            • 0 Attachment
              > One of your basic assumtions is flawed. It is that the growth in the
              > availability of
              > bandwidth with outpace growth in storage capacity.

              Hmm. I'm not sure that what we are farming out to the internet is storage.
              Rather than dumb hard drive capacity, it is decisions and expertise.

              For example, I think the reason we don't just buy all of CPAN on a cdrom every
              so often is that it is changing too fast for a snapshot to be anywhere near as
              good as the most recent uploaded version. The value is not that CPAN stores it,
              it is that CPAN tracks it.

              > I don't see any indication that Napster or P2P technology is or could
              > interfere with
              > AOL/TimeWarner and the business model.

              Ok. Have AOL re-start Gnutella with full corporate backing and see how loud
              TimeWarner yells.

              > Napster type technology may be
              > the earliest
              > indicator that information distribution will not be a viable business
              > for much
              > longer. The problem is that AOL/TimeWarner is not exclusively or even
              > primarily a
              > distribution business. They are in the business of producing material
              > on behalf of
              > sponsors(advertisers), and they are in the business of providing a
              > distribution
              > medium on behalf of the distribution businesses(ie: cable networks).

              lovely points.

              > AOL and Cable networks and other business in the distribution will
              > probably suffer
              > as the availibility of bandwidth and storage capacity increases, but the core
              > businesses will not suffer.
              >
              > It is necessary to face the fact that only a very large centralized
              > production
              > company can produce entertainment of the quality and quantity we now
              > expect.

              I don't completely agree - weblogs are both compelling, popular and
              decentralized - but I see your point. The big content makers dominate for good
              reasons. On the other hand, the good reasons are mainly about economies of
              scale driving competing creators out of the distribution channel.

              - Lucas

              (p.s. sorry about semi-random line breaks above. ...have recently been forced
              to switch to Outlook, which does hell to replies).
            • frank black
              ... storage. ... Excellent point, but I think in the long run that the raw storage capacity available to everyone will reduce the internet to something little
              Message 6 of 9 , Aug 28, 2000
              • 0 Attachment
                --- In decentralization@egroups.com, "Lucas Gonze" <lucas@w...> wrote:
                > > One of your basic assumtions is flawed. It is that the growth in
                >
                > Hmm. I'm not sure that what we are farming out to the internet is
                storage.
                > Rather than dumb hard drive capacity, it is decisions and expertise.
                >
                > For example, I think the reason we don't just buy all of CPAN on a
                >....snip...
                >CPAN stores it, > it is that CPAN tracks it.

                Excellent point, but I think in the long run that the raw storage
                capacity available to everyone will reduce the internet to something
                little more complex than a usenet feed. Please forgive the
                oversimplification. If the storage in our homes is so massive that we
                can simple recieve the a constant stream of DIFFs of everything on
                the internet to archive which are applied on the fly as we use
                content
                we get the value of the CPAN CD in addition to the value of CPANs
                expertise and labor.

                >
                > Ok. Have AOL re-start Gnutella with full corporate backing and see
                how loud
                > TimeWarner yells.

                I think TimeWarner would be quite happy if they realized how they
                could profit if they maintained control of Gnutella from the outset.
                The only way the banner ad profit model could succeed is on this
                scope, and if they maintained the web interface to the network,
                imagine the numbers of ad impressions. BTW, Im convinced a web
                interface to gnutella's search is the way to go, considering
                freeriders and all.

                > > company can produce entertainment of the quality and quantity we
                now > > expect.
                >
                > I don't completely agree - weblogs are both compelling, popular and
                > decentralized - but I see your point. The big content makers
                dominate for good
                > reasons. On the other hand, the good reasons are mainly about
                economies of
                > scale driving competing creators out of the distribution channel.

                I don't completely agree with my last point either. I find web radio
                compelling, just like i've always found shortwave radio, ham radio,
                zines, and pirate radio compelling. Some people find webcams
                compelling. In some respects the internet is making the promise of
                cable television a reality. But I disagree with the reason. I think
                it
                is the web acting as a centralizing force that allows competition in
                markets a business could not afford to address, rather than big
                business freezing out competition.
              • Lucas Gonze
                ... I d like add one thing to this - p2p provides the ability have a group of small iron providers (a bunch of home PCs, for example) work together to provide
                Message 7 of 9 , Aug 29, 2000
                • 0 Attachment
                  > Where P2P is important to me is the (a) groupware, common
                  > bookmarking
                  > aspect where I can find my demographic and run with similar people, and
                  > (b) it provides a visibility channel that the web failed to provide: a small
                  > news shop can now compete with a big shop on a reputation basis to sell
                  > the same news, or product, and have a dedicated following from over the world
                  > and be viable.

                  I'd like add one thing to this - p2p provides the ability have a group of small
                  iron providers (a bunch of home PCs, for example) work together to provide a
                  resource. Where operating a web server is 24-7 or nothing, which pretty much
                  makes home web servers impractical, any 20 little guys can get together to
                  mirror a site and be pretty sure that one of them is up and running all the
                  time. ...the upshot of which is to say that functional routing allows for
                  clumps of little guys to compete with the big guys.

                  - Lucas
                • Lucas Gonze
                  ... I love this vision but think it is far off enough that, for the moment, pretending it isn t there will help get work done. ... not so hard, and maybe it is
                  Message 8 of 9 , Aug 29, 2000
                  • 0 Attachment
                    > Excellent point, but I think in the long run that the raw storage
                    > capacity available to everyone will reduce the internet to something
                    > little more complex than a usenet feed. Please forgive the
                    > oversimplification.

                    I love this vision but think it is far off enough that, for the moment,
                    pretending it isn't there will help get work done.

                    > BTW, Im convinced a web
                    > interface to gnutella's search is the way to go, considering
                    > freeriders and all.

                    not so hard, and maybe it is interesting that is a vote for the
                    fractional-horsepower-http faction, since all the servants would need mini-web
                    servers.

                    - Lucas
                  • mark
                    ... On gnutella, all members are a mini web server. Gnutella serves files using http get commands on port 80. I meant to say the access point to the network
                    Message 9 of 9 , Aug 30, 2000
                    • 0 Attachment
                      > > BTW, Im convinced a web
                      > > interface to gnutella's search is the way to go, considering
                      > > freeriders and all.
                      >
                      > not so hard, and maybe it is interesting that is a vote for the
                      > fractional-horsepower-http faction, since all the servants would need mini-web
                      > servers.

                      On gnutella, all members are a mini web server. Gnutella serves files using http get
                      commands on port 80. I meant to say the access point to the network should be a web
                      interface for most people. Surfy gnutella is a simple one that I use,
                      http://www.surfy.com/page/nut.php3. There are manifold benefits to this approach. It
                      is easy to extend, the user always has the most up to date version of the client, the
                      user has no network overhead. Going back to usenet, its the difference between have a
                      usenet feed and having an account on a nntp server. The intermediate server is the
                      better choice for 90% of users.
                    Your message has been successfully submitted and would be delivered to recipients shortly.