World Wide Web at

1
World Wide Web at

• 1999 Computing Advisory Committee Review
• Ruth McDunnWeb Information Manager
• Dennis Wisinski
• The Applications Group (TAG) Manager

2
Topics

• A bit of web history
• Review progress against our strategic plan
• Vision
• Goals
• Progress
• Needs
• Summary
• Resource Issues

3
Our Role in the WWW

• In December 1991, SLAC installed the first web
  server outside of Europe
  
• SLACs early web presence
• Reflects in established, legacy procedures
• Some good, some bad, difficult to change
• Has led to a very large and tangled web
• Couldnt plan for the technology advancements
• We cant start from scratch like those entering
  the web today

4
How Big Are We?

• As of June 27, 1999
• Central SLAC web
• Unix and most of the NT servers
• 415,162 documents and 1,010,688 URLs
• BaBar web
• 15,068 documents and 26,354 URLs
• Figures do not include
• www-user and defined development spaces
• slaconly and other areas with access controls

5
Main Unix ServerCurrent Statistics

• June 15-25, 1999
• Average total hits
• 132,413
• 149,841 (weekdays)
• Total KB transferred
• 1,607,429
• 1,802,599 (weekdays)

6
Vision

• To provide a ubiquitous and efficient
  communication system enabling SLACs internal and
  external community to share information. In turn,
  allowing SLAC to perform collaborative,
  scientific, and technical work without
  constraints imposed by locations, hardware,
  software platforms, or unwarranted administrative
  obstacles.
• Strategic Plan for Computing
• Section 18, World Wide Web at SLAC

7
Goals

• Provide highly available access to information
• Provide high quality and accurate information
• Provide easy-to-use access to information
• Draw in WWW support and expertise from across the
  lab
  
• Provide appropriately secure access to information

8
Provide Highly Available Access to Information

• Central web servers
• Unix (Netscape/Apache) and NT (Microsoft IIS4)
• Uninterruptible power supply
• Nightly incremental backups
• Network monitoring every 15 minutes, daily
  heartbeat
  
• Long term logs show very high availability
• Slowdowns usually involve local and remote robots
  indexing the site, script or AFS problems, or
  excessive network traffic

9
Provide Highly Available Access to Information

• Need
• Move problem services (such as ASP scripts) to
  their own server
  
• Establish fault-tolerance, mirroring, and load
  balancing for Unix web severs and clustering
  techniques for the NT servers
  
• Develop scripts to monitor and possibly restart
  critical services, as necessary
  

10
Provide High Quality and Accurate Information

• Many aspects of web policy, procedures, and
  resources are documented and advertised
  
• Standard software with defined support
• FrontPage, FTP, browsers
• Web Information Manager appointed
• WebAnalyzer Professional purchased
• Web interface to retrieve PeopleSoft data
• Purchasing, Property Control, Stores, Financials

11
Provide High Quality and Accurate Information

• Need
• Continued development of web interface to
  PeopleSoft and other institutional data
  
• Input data, change data, and query
• More author friendly tools
• Considering FrontPage extensions and ASP on Unix
  and PERL CGI for NT servers
  
• QA program to spot review web sites
• Link analysis, outdated materials, meet page
• requirements

12
Provide High Quality and Accurate Information

• Need
• Better log analysis tools
• Unix logs range from 1040 M daily and growing

• Consistent format on all six real and virtual
  servers (production space)

13
Provide Easy-to-use Access to Information

• Year long meetings to revise home pages
• Redesign implemented August 1998
• Phonebook search
• Show Index script
• Dynamically generates alphabetical list of links
  on a page
  
• InfoSeek search engine/indexer
• SLAC-specific index started late 1998
• Web space requests go through WIM

14
Provide Easy-to-use Access to Information

• Need
• More active review of current products
• Web interface to database
• Server technologies
• Reorganization and maintenance of lower level
  pages
  

15
Draw in WWW Support and Expertise From Across the
Lab

• Central Support -- 3.8 FTE
• Web Information Manager (.5 FTE)
• Infrastructure in TAG (2.3 FTE)
• Unix Web Admin (1 FTE)
• NT Web Admin (1 FTE)
• Direction, policy, programming (admin),
  security(.3 FTE)
  
• Content in TAG (1 FTE)
• Documentation/computing web maintenance (1 FTE,
  vacant)

16
Draw in WWW Support and Expertise From Across the
Lab

• Distributed Support
• Web Coordinating Committee (12)
• Web Technical Committee (33, informal)
• Welcome Page Committee (4)
• Web Support Coordinators (40)
• Computer Coordinating Committee (13)
• 102 people on these combined lists

17
Web Management
Laboratory Director
SCS
Web Support
Coordinators
Chair - J. Masek
Divisional Web
Management
18
Distributed Web Support (71 people)
19
Draw in WWW Support and Expertise From Across the
Lab

• Presentations
• Searching and Being Searched, Java Overview,
  Introduction to JavaScript, About the New SLAC
  Home Pages, Web Management at SLAC, FrontPage at
  SLAC, Server Tricks, Cascading Style Sheets
• Training Courses
• So You Want (Need) to Be a Web Author, FrontPage
  98, Introduction, Forms in FrontPage 98

20
Draw in WWW Support and Expertise From Across the
Lab

• Need
• Appointment of WSCs to remaining departments
• Immediate supervisor support for WSC and web
  author activities
  
• Full day classes
• Meeting participation
• More staff
• Develop and provide training
• Develop materials related to our science
• Review commercial products

21
Provide Appropriately Secure Access to Information

• Security issues driven by PeopleSoft project
• Needed encrypted transport of passwords and data
• SLAC only, group only, password protected access
  via SSL on NT web server
  
• Computer Security Officer/TAG
• Computer security and web security policies web
  pages
  
• Servers updated as fixes are issued
• Server configuration changes restricted to a
  short list
  
• Eliminated FTP access to NT web servers
• Blocked non-approved port 80 web servers

22
Provide Appropriately Secure Access to Information

• CGI wrapper
• CGI script authors are tutored in security
  issues
  
• Only Bebo, Dennis, or Les can add screened
  scripts to the wrapper
  
• Simple checking on the input to the user's CGI
  script
  
• Allows authorized UNIX commands
• Specifies the server (test or production) it runs
  on and if it is slaconly

23
Provide Appropriately Secure Access to Information

• Need
• SSL on Unix (testing Apache)
• Move the Unix and NT test servers to separate
  physical machines
  
• Scripts to check and correct insecure
  directories
  
• Explore JAVA servlets and CORBA technologies
• Reduce dependence on CGI for legacy systems
• Implement method to detect web server compromise

24
Summary

• Needs
• Server Issues
• Content Analysis
• Content Interface
• Content Development
• Implementation
• Human Resources
• Other Resources

25
Server IssuesSummary

• Move problem services to their own machine
• Move test servers to separate physical machines
• Establish server redundancy, mirroring,
  balancing
  
• Establish SSL capabilities on the Unix server
• Develop scripts to monitor (restart) critical
  services
  
• Develop scripts to check and correct insecure
  directories
  
• Explore alternatives to CGI and evaluate new web
  technologies for strategic use applications

26
Content AnalysisSummary

• Purchase a robust log analysis tool
• Collect log data in a consistent format on all
  servers
  
• Create a QA program to spot review selected web
  sites within the SLAC web
  
• Obtain appointments of WSCs to remaining
  departments (who have, should have, or want a web
  presence)

27
Content InterfaceSummary

• Continued development of web interface to
  PeopleSoft and other institutional data
  
• Develop interface to non-institutional data
• Evaluate FrontPage extensions and ASP for Unix,
  PERL CGI for NT servers, Office and FrontPage 2000

28
Content Development Summary

• Perform more reviews of current web authoring and
  website management products
  
• Assist authors with reorganization and
  maintenance of lower level pages
  
• Request immediate supervisor support for WSC and
  web authoring activities
  
• Provide/develop more and deeper author training
• Develop more web pages to present our science to
  the public

29
Human Resource Needs

• Infrastructure (TAG)
• 1 to 2 FTE - Backup administrator, web tool and
  technology evaluation
  
• Intellectual Content/Structure Support
• 1 FTE - Web site design and (re)organization,
  training, log analysis, content review below the
  top level
  
• .5 FTE - Writer to explain our science to the
  public (not just for the web)

30
Other Resource Needs

• Cisco Local Director
• Router to provide server redundancy, mirroring,
  and load balancing for Unix servers
  
• Need to find an equivalent system for NT servers
• Continuing upgrade of server hardware
• Net.Analysis
• Site traffic analysis tool that works across all
  servers
  
• Looks at packet traffic, not web logs

31
More Information

• Snapshot review of the Strategic Plan on
  Computing, Section 18
  
• http//www.slac.stanford.edu/slac/www/wwwcc/wim/1
  999review.html
  
• Questions