200 likes | 386 Views
The Dance of Co-Opetition. Dave Crocker Brandenburg Consulting MY: +60 (19) 3299 445 www.brandenburg.com US: +1 (408) 426 9827 dcrocker@brandenburg.com Fax: +1 (408) 273 6464. Dave Crocker Internet since 1972 Email, EDI, Fax, ... TCP/IP, Net mgmt Standards Development
E N D
The Dance of Co-Opetition Dave Crocker Brandenburg Consulting MY: +60 (19) 3299 445 www.brandenburg.com US: +1 (408) 426 9827 dcrocker@brandenburg.com Fax: +1 (408) 273 6464
Dave Crocker Internet since 1972 Email, EDI, Fax, ... TCP/IP, Net mgmt Standards Development Product, service MCI Mail, DEC, SGI Startups Consulting Services Internet business planning Product, system design Technical audits Standards track/contribute Brandenburg Consulting
The Internet world today • Useful (boring) messaging • Web excitement • Multi-media • Marketing • Information access • Huge installed base • Easy entry • Explosive growth Good News
The Internet world today • Limited bandwidth, • Variable delay • Information searching difficult • Confusing, inadequate security • Inconsistent management • Proprietary extensions Bad news
Overview • Open vs. proprietary • Networking requires open • Disturbing trends? • Competitive pressures • Fiefdoms vs. community • Core vs. edges • Infrastructure takes time
Open vs. proprietary • Proprietary • Control • Focus • And timeliness • Open • Multiple vendors • Broad review • Generality
The meaning of “open” Publication:Any may read & implement Ownership:Group control of specs Development:Broad participation
Internetworking requires open • Casual interaction • Without prior arrangement • Participants must support same set of capabilities • Fragile basis • Deviation by any components prevents interoperability • Standards “based” isn’t good enough
Styles of use • Receiver pull • Interactive sessions • Individual, foreground refinement • Sender push • Messaging • Bulk, background distribution (Mark Smith, Intel)
Upper vs. lower layers • Open transport / Proprietary applications? • Still requires prior arrangement • Still requires multiple apps for same task • Explosion of user complexity • Increased price
Quicker to market Carefully tailored to vendor need Creator benefits Non-interoperability Different package for every function Inadequate public review Competitive pressures
Core vs. edges... My object Channel Object Secure Web Security EMail Security My object Web Server MTA EMail Web FTP Web Server MTA EMail Secure Secure My object My object My object My object
Core • Infrastructure • Support along entire path • Adoption delay • Operation fragility/dependence • No central control • Time before useful / popular • Decade
IPv6 • Began with simple goal • Increase address space • Became design by committee • Should have deployed 3 years ago • Lucky to get any installed base by 2000
Edges • Any two hosts • Instantaneous utility • No special infrastructure benefits • Plus • Minus • Time before useful / popular • Year / half-decade
Intra- vs. Inter-nets? • Intranets • Move to ISP administration style • WAN lines usually congested • Internets • Virtual corporations need public facilities
Integration • System operators • Hate extra boxes • Users • Hate extra applications • Except when they love them
Facts of life • Real-time Global Internet • 5-10 years, minimum • High-bandwidth to global users • 5-10 years, minimum
Cliches to live by • Customers buy solutions • A product that solves three problems • Is better than one that solves only one
Fiefdoms vs. community? • Vendor initiatives • Market lead • Folded into public standards • Open access • Open enhancement It all depends on market demand.