Milestone No. 6 and 7 Completed
Date: 30
Novemberptember 2001
1.0 Deliverables
There were 2 major
objectives in Milestones 6 and 3 objectives in Milestone 7. Since the first 2 objectives of each of
Milestones 6 and 7 were a continuation of the same activity, we will comment on
each of these objectives together:
1.1 Operations group completes the Beta
version (of phase 2) and the Final version (of phase 1) Management Software Done
1.2 Evaluation committee assesses and
approves the Beta and Final versions of the software, and suggests final
refinements Done
OVERVIEW
While proceeding to outline what was done in Milestones 6 and 7, we would also like to review what has been accomplished to date, and suggest the major activities for the final school year. The review will be in two parts: Hardware and Software, with software discussion covering the SchoolWeb application and Cache monitoring. Much of the specific detail required for this Report will be provided in this summary.
1. Hardware
Regarding the hardware portion of the SchoolWeb Project, 19 school servers, with Version 1 software along with our headend was designed, built and installed in the first 9 – 10 months of the Project. The hardware featured a Linux operating system, with dual SCSI drives, standby power and redundancy for high reliability. The hardware has performed very well, with only two component failures over the first year of the project. There was no major downtime or loss of operation recorded at any school over the active time period. Documentation for ‘Administration’ of the server has been provided and upgraded as new features were added over the past 9 months.
Constant work is required to handle the changes in District configurations, PCs vs. Mac installations, firewalls, intermediate proxy units, router port configurations, IP/Gateway changes, student names/password changes, interoperation with other networks (Novell, etc.), and interoperation with network tools (Deepfreeze, etc.)
At the start of this school year a short review was sent to each school to make sure that the admin changes were being made. (For more information on these topics, See “To all SchoolWeb Administrators” – 6 pages)
In all cases, quick solutions to these normal problems have been found and implemented. This process continues as networks in School Districts continue to change.
By March 2002, a final review of the hardware and its operation will be made; and recommendations will be put forward on any upgrades or changes to be incorporated before June 2002.
Software
The SchoolWeb software is discussed in 2 parts:
1. The SchoolWeb Librarian application, (including dynamic and policy-based caching, ordering sites, reviewing and changing sites, and searching and retrieving files)
2. Cache performance monitoring
1. The SchoolWeb Librarian is the recent name given to the second phase of Software Development. The first phase of software concentrated on getting dynamic and policy-based caching to work via line and Broadcast (using date-in-video) means. This version was placed on each of the installed SchoolWeb servers as the units were being installed, and the software updated as new revisions and features were added. By March 2001 this version was operating satisfactorily, and was considered to be final for Phase 1.
The second phase of software development attempted, not only to provide multiple media files and manipulate/manage them, but also to streamline the methods for caching them, making the access process more user friendly for Teachers and Librarians, and placing a search engine on the server to ensure that the Policy-Based Cached (PBC) files could easily be found and used by Teachers and students. Early version of some of the programming for Phase 2 was completed in April, May and June 2001, and a preliminary version of the software was cobbled together for demonstration at an Evaluation Committee meeting on June 13th in Vancouver. (For more information see “Minutes of A Meeting of SchoolWeb Partners and SchoolWeb Evaluation Committee Members” – 3 pages). Full meeting information and PP Presentations are available on request.
The Evaluation Committee made a number of suggestions, many of which were taken into consideration when the first working version of the software was completed in early September and placed on all SchoolWeb Servers.
Version 2 of the software had a totally new look and feel. It starts from a SchoolWeb Administration page, and provides:
-
A simple method and dialog page to “Order a site” for
policy-based caching with lots of features such as either http or ftp URL
input, variable recursion depth capability, Dewey Classification drop-down
boxes, subject listings, annotationsgrades, type of site,
notes, caching period start and expiry dates, and selectable file
website
refresh rates.
- A full listing of all requested sites that are being pushed to the cache, and the ability to see/change the details of the site, modify the request or delete the site.
-
The delete feature is currently limited to an “admin” user
(with a password) and Advanced Interactive Canada, Inc control’s this user account.who
this user will be. At this
time, the delete feature only deletes the websites at the headend server, but
in the next short while, the delete feature will be extended to also delete the
complete website from the selected Schools’
servers. We are currently upgrading the
ability to send the Dewey Classification, Grade Information, Lesson
Type of
Site, and
Notes,
along with
the and
Expiry Date with the transmitted files to the School Web.
SchoolWeb
servers.
-
Once the files are transmitted to the School Web servers they
are added to the Squid Cache immediately.
A copy of the new files is kept in a temporary directory. Over night, all the files in the temporary
directory are analyzed and summaries of each of the files in the temporary
directory are created. The summary
files are then added to the School Web SchoolWeb Librarian
Data Base. The database is then indexed
and catalogue. Currently, the data can
be searched by:
- Full-Text
- Keywords
- Title
- Description
- Author
Once the Dewey Classification, Lesson Type, Grade, and Notes is transmitted with each of the files to the schools. All these parameters will be added as search elements.
Three ways (
simple, advanced, expertSIMPLE, ADVANCED,
EXPERT) of accessing a customized search engine under the School
Web SchoolWeb
Librarian tab are provided to allow quick and
convenient methods for students and teachers to search the Policy-Based Cache
(PBC) files.
-
SIMPLE Search
-
SIMPLE search: results
-
ADVANCED Search
- ADVANCED search: results
- EXPERT SEARCH
-
EXPERT search: result
On the same page, is a link showing all the sites cached, with title and annotations that are in the PBC.
At the same time that the new software was placed on all SchoolWeb Servers, instructions were sent out to all Administrators, Teachers and Librarians in Schools using the SchoolWeb Caching Servers. (See attached ‘Instructions to all Administrators, Teachers and Librarians using the SchoolWeb Caching Server – 12 pages). These instructions explain to Teachers and Librarians, how easy it is to order sites and make sure they are in the PCB cache.
Over the coming months, our activities will include:
1. Monthly upgrades to the software, as new experience is gained and new features added.
2. An increase in training activities for each of the schools, to directly contact Teachers and Teacher-Librarians, to demonstrate the features and advantages of PBC, and obtain feedback.
3. Developing ties into the BC Provincial Portal, to cache (PCB style) important learning outcomes and lesson plans for students and Teachers
4. Exploring and implementing additional recommendations of the Evaluation Committee from the June 13th meeting and other regular contacts.
5. Planning for the continuation of the SchoolWeb Program beyond June, 2002
2. Cache Performance
The first version of monitoring the use of the cache, provided statistics and tables on the number of Hits and downloaded files (MB) from the Internet and from the cache. These Hits/MB and times were compared in table and chart form. The chart diagram was considered by users to be most useful, and they asked that that service be continued. However, at that time, we did not distinguish between files placed in the cache by PBC vs. any other means.
In Milestones 6 and 7, we moved to incorporate PBC files into the comparison, and added this feature on a first trial basis, in the first release of Phase 2 software.
Below is a screen showing the summary results for the Month of October 2001.
Added to the previous stats were the following columns of information:
Number of client machines connected
Number of MB transferred from the SchoolWeb cache using PCB*
Average time to transfer 1 MB from the SchoolWeb cache using PCB*
Average time to transfer 1 MB from Internet without using cache
Speed improvement (Internet to SchoolWeb cache)
Speed Improvement (Internet to PBC*)
Speed Improvement (not using cache vs.
dynamic caching)
Speed improvement (Not using cache vs. PBC* cache)
Although we have not finalized this feature fully, this additional statistic should result in a much more accurate view of the speed advantages of PBC. As the use of PBC builds up, we will be able to clearly differentiate the speed differences of files that must be retrieved from the Internet, vs. files that must be
checked by the host internet site, vs. files that are already in the cache and current. Early results are encouraging.
The last
column shows the speed increase ratio for using Policy Based Cache (PBC)
compared to its normal Internet connection and using a benchmark figure for
downloading 1 MB of information. The PBC/Internet ratio for schools with low
speed connections (64Kb/sec) is very high e.g. Big Eddy is 76.7:1 and
for a large high school e.g. Burnaby South with a high speed connection and 500
computers it is still a very respectable ratio of 11.27:1; in the case
of Burnaby South this translates into effectively boosting its already high
speed connection by 11X. The statistics further show that the delivery
of 1MB of PBC data from the Cache to the client machine is relatively
independent of the Internet connection speed at the school. At Big Eddy the delivery of 1MB of PBC data
from the cache took 19 seconds and at Burnaby South it took 37 seconds. The delivery of the PBC data from the cache
is more a factor of the LAN speed and LAN congestion than on the Internet
connection speed at the school.
Our next activities in this area will be to monitor the results and recheck our methods to ensure that we are getting a representative view of what is actually happening. Upgrades and improvements, along with a more automated process are yet to be established. At the moment, manual intervention is required.
In the final report, these statistics will be used to show the relative speeds using the 3 methods of delivering Internet files to schools, and will also take into consideration, the speed of the line connection. (For a view of the Stats go to “Cache performance” tab on the School Administration page)
HIGH RELIABILITY DATA WITH COMPRESSION
SchoolWeb Software V1 added bundled FEC coding to the data transmitted to the schools to improve error correction. The bundled FEC code effectively eliminates the missed data lines by replacing them with the redundant data contained in the bundled FEC code. The bit error rate with the bundled FEC is now better than 10-9. We also added file compression to make our transmissions more efficient.
More recently, we had to modify the priority of the receiver
program to be higher so that packets were not missed due to the SchoolWeb
server CPU being
busy and not scheduling giving the receiving program. By increasing the
receiving programs priority we have almost eliminated this problem. Linux not being a real-time system we can not guarantee
that a program will be schedule. To eliminate this problem we should consider
including the receiver software to run within the device driver code rather than as
a sub-process. This will guarantee that packets are not missed even when the CPU is
busy since
the device driver code is run on hardware interrupts and not on Round Robin scheduling.
We upgraded the
software to modify the expiry time on the files received before adding them to
the Squid Cache. All Policy Based files
have an expiry time. By having an
expiry time the Squid Cache does not check if the file is fresh before it
serves it as long as the expiry time is after the current system time. This provides a marked improvement in
delivery speed of the files from the cache to the Client machine. In effect, the delivery times
are controller by the LAN speeds rather then the Internet Connection speeds.
This software needs
to be further improved to also send the following information from the Headend
to each of the SchoolWeb Servers:
1. Expiry Time
2.
File
Type as specifically received from the httpd server rather than using the file extension
to figure out the file type.
3.
Dewey
Decimal Classification
4.
Type of Site
5.
Grades
6.
Notes
Another feature that
needs to be completed is the deletion of files from each of the
servers
once a requested URL is deleted before its expiry time.
Files are also added to a temporary storage directory for processing at night. Each night all the files received are added to the SchoolWeb Librarian Data Base. The Squid Cache and SchoolWeb Librarian Database need to be better synchronized so that all the files active in the Squid Cache are also available in the SchoolWeb Librarian Data Base. Each night, the SchoolWeb Librarian Data Base is re-indexed after all the new files, received in the previous 24 hours, are added to the SchoolWeb Librarian Data Base. The switch over of the database is done within a few seconds.
WEBSITE FILTERING
When SchoolWeb servers were initially installed we were
relying exclusively on PLNet’s website filtering to prevent restricted websites
from being accessed through SchoolWeb servers.
At a few of the schools, access to the Internet is not through the PLNet
but directly to the Internet via the local Internet Service Provider
(ISP). This was causing concern to
these schools since some restricted sites could be accesses by students. Therefore, we enabled a website filter in
the sSquid cCache along
with a list of restricted websites.
School administrators are given the capability of adding restricted
sites and/or legitimate websites to the appropriate lists to suit their needs.
We also provide the website filtering utilizing the DNS (Domain Name Service) based filtering provided by the BC Government. The two DNS servers we use for this service are located at:
DNS1.GOV.BC.CA IP: 192.75.26.15
DNS2.GOV.BC.CA IP: 142.22.250.77
Both services are available on our SchoolWeb servers. The school administrator at each school can
decide which service(s) they would like to utilize on the SchoolWeb server
located at their school.
Some of the schools had previously purchased Web Filtering Proxy
servers. At these Schools the
SchoolWeb server(s) were made siblings to the Proxy Servers located at the District Office. This allowed the Web Filter Proxy Servers to
continue to provide first level filtering
of websites for the SchoolWeb servers. The only disadvantage of this
arrangement is that the SchoolWeb access speeds are slightly higher due to having to
check with the
Proxy server that were is usually located at the District Office. This ese servers
often serves a large number of schools, since one Proxy server is used for all
the schools in the School District. A
distributed or edge Proxy/Cache server would provide substantially higher speeds that a cascaded
Proxy/Cache server chain.
AUTOMATED STATICTICS COLLECTION SYSYEM
Up to now, a completely automated system of collecting
SchoolWeb statistics for each of the SchoolWeb servers has been working
well. Each weekday night the Squid log
files are rotated and the old files are sent to AIC’s main Broadcast server
via e-mail for analysis. On receipt of
the files, the log files are analyzed with an analysis tool and the summary of
the analysis is posted on a Member-only website where the summary of the
statistics for each of the schools is maintained. This information is available for the SchoolWeb server administrators/teachers
to review. On a monthly basis, a
complete month’s statistics are analyzed and a summary of the information is
made available on the website. (i.e. http://www.advancedinteractive.com/secure)
as part of our new SchoolWeb page.
A
new feature recently added is that, from the access.log files received from each of the
Schools, a
MYSQL database is populated that only includes files added to the School’s
cache that are Policy Based Cache (PBC) Ordered Files. The same access.log files are also analyzed
to check if any of the PBC files added to the Squid Cache have been requested. The statistics for PBC files are kept separately and reported
separately in the monthly statistics.
The results for the Months of October and November 2001 show that PBC
files get delivered faster than non-PBC files.
Furthermore, PBC file delivery is substantially constant
regardless of the speed of the Internet connection at the School.
A summary report of
the October 2001 Schoolweb server statistics is show below. As can be seen, schools with low
speed lines see much more speed improvement in files that have been
placed in the cache using PBC, than without policy-based caching.
|
|
School |
Number of clients |
Mbytes transferred from Internet |
Mbytes transferred from SchoolWeb cache |
Mbytes transferred from SchoolWeb cache using
PBC* |
Average time to transfer 1MB from Internet |
Average time to transfer 1MB from SchoolWeb cache |
Average time to transfer 1MB from SchoolWeb cache
using PBC* |
Average time to transfer 1MB from Internet
without using cache |
Speed improvement (Internet to SchoolWeb cache) |
Speed improvement (Internet to PBC*) |
Speed improvement (not using cache at all to
caching) |
Speed improvement (not using cache at all to
PBC*) |
|
bee |
Big Eddy Elementary |
26 |
110.45 |
43.44 |
1.40 |
0:17:03 |
0:01:49 |
0:00:19 |
0:23:45 |
9.40 |
55.07 |
13.09 |
76.72 |
|
rupert |
Fort Rupert
Elementary |
18 |
1,376.59 |
1,635.63 |
1.46 |
0:02:22 |
0:00:16 |
0:00:17 |
0:05:10 |
9.01 |
8.32 |
19.71 |
18.21 |
|
rss |
Revelstoke
Secondary |
38 |
3,964.81 |
4,334.09 |
8.45 |
0:03:00 |
0:00:21 |
0:00:12 |
0:06:17 |
8.59 |
14.72 |
17.99 |
30.81 |
|
nss |
Nakusp Secondary |
19 |
1,347.27 |
337.32 |
0.72 |
0:02:52 |
0:00:28 |
0:00:09 |
0:03:35 |
6.25 |
19.02 |
7.82 |
23.78 |
|
hutton |
John A Hutton
Elementary |
33 |
452.89 |
367.78 |
0.49 |
0:02:37 |
0:00:26 |
0:00:12 |
0:04:45 |
6.01 |
13.43 |
10.89 |
24.33 |
|
westview |
WestView Secondary |
100 |
6,416.26 |
3,987.21 |
11.67 |
0:02:23 |
0:00:29 |
0:00:12 |
0:03:52 |
4.95 |
11.44 |
8.03 |
18.55 |
|
hardy |
Port Hardy District
Office |
124 |
4,315.68 |
2,386.02 |
16.17 |
0:03:54 |
0:00:57 |
0:00:20 |
0:06:03 |
4.09 |
11.91 |
6.36 |
18.49 |
|
bss |
Burnaby South
Secondary |
201 |
16,780.07 |
6,760.00 |
30.92 |
0:04:56 |
0:01:18 |
0:00:37 |
0:06:56 |
3.79 |
8.03 |
5.31 |
11.27 |
|
gladstone |
Gladstone Secondary |
39 |
3,057.29 |
2,212.71 |
2.05 |
0:07:40 |
0:02:03 |
0:01:03 |
0:13:13 |
3.73 |
7.25 |
6.43 |
12.50 |
|
rcs |
Robson Community |
10 |
11.19 |
2.35 |
- |
0:06:19 |
0:02:06 |
|
0:07:39 |
3.01 |
|
3.64 |
|
|
bns |
Burnaby North
Secondary |
165 |
1,535.46 |
935.57 |
5.61 |
0:05:38 |
0:01:55 |
0:01:13 |
0:09:04 |
2.93 |
4.63 |
4.71 |
7.46 |
|
fulford |
Fulford Elementary |
40 |
38.76 |
28.24 |
0.39 |
0:19:42 |
0:09:17 |
0:00:41 |
0:34:04 |
2.12 |
28.82 |
3.67 |
49.83 |
|
proxy1 |
Fort Nelson
Secondary |
4 |
16.99 |
1.69 |
- |
0:05:50 |
0:02:59 |
|
0:06:25 |
1.96 |
|
2.15 |
|
|
topham |
Topham Elementary |
20 |
18.84 |
1.52 |
- |
0:01:59 |
0:01:05 |
|
0:02:08 |
1.82 |
|
1.97 |
|
|
bcs |
Boundary Central
Secondary |
19 |
164.31 |
38.64 |
2.53 |
0:11:26 |
0:10:34 |
0:01:01 |
0:14:07 |
1.08 |
11.32 |
1.34 |
13.98 |
POLICY BASED WEBSITE ORDERING
Improvements to the SchoolWeb ‘Order-a-Site” page have been discussed earlier in this Report. In addition to the previous comments, it should be remembered that our software also builds a data-base of the Dublin Core Metatags. These tags are useful in enhancing the search capability.
Since our last Report, we have devised a way to identify objects that are policy-based, compared to dynamically requested objects in the cache. Work in this area is continuing and other techniques will be tried to further optimize the Broadcast Cache Server. We are just starting to obtain some results, which will indicate how the cache speeds have been impacted due to PBC.
Our early results indicate that there is
a substantial improvement when PBC files are delivered to Client Machines, independent of the
School’s Internet Connection speed. To
further expand on this capability it was felt that we should try to encourage students and
Teachers to
concentrate on websites in the PBC rather than on the whole Internet. We have been encouraging schools to use the
SchoolWeb Librarian for all searches first then using the other search engines available on
the Internet
with an automatic search link to a few of the more popular sites.
An explanation of the process has been added to the SchoolWeb Client Administration/Operator’s Guide, a publication that assists the school to operate the SchoolWeb server
RECENT EVALUATION
COMMITTEE COMMENTS
On June 13th, 2001 a meeting of the
Evaluation Committee was held in Vancouver at the BCIT Downtown Campus
location. An Agenda was prepared and
distributed in advance, along with a list of Topics to be discussed, outlines
of PowerPoint presentations, and examples for searching the SchoolWeb Librarian,
to stimulate ideas. The meeting was
attended by representatives from both elementary and secondary schools,
CANARIE, BC-MOE, SchoolWeb Systems Inc. (our marketing partner), and other
invited guests.
The results of the meeting were very helpful
in guiding our development over the past 3 months, and confirmed some of the
directions we had taken from the prior 3 months.
Rather than summarize the results again, the
Agenda, Minutes of the meeting, List of Participants, Topics, and Examples,
have been sent to all evaluation Committee members, and are available from
Harry Davis, CEO Advanced Interactive Canada Inc. on request.
In addition, other positive feedback has also
been received from Evaluation Committee Members in response to mailings such as
how to “Order-a Site”,
1.3 AIC completes tests for improvements
in delivery of files to students at home. Deferred
At a meeting with David Sutherland, Peter Marshall and Colleen Bretzlaff of CANARIE, it was suggested that we place a low priority on work in this area of activity, since it did not have such a high impact on the caching experiment in schools. This did not mean that we would abandon this activity altogether, but it would only be done if extra time was available.
2.0 Schedule
No substantive changes have been made to the schedule that was last submitted as Gantt Chart V7, revised to 31 March 2001. the only potential change would be made if the Project ws extended from March 31, 2002 to June 30, 2002.
3.0 Budget
3.1 The following are comments on the budget status of the project (see also the budget/actual comparison and refer to the Worksheets and Milestone Budget Summary, revised 01 April 2001)
1. Direct Labor
The costs incurred for Direct Labor for this Milestone period are higher than budgeted, due to the inclusion of costs previously disallowed because payment was not made during the milestone period. However, the accumulated actual have now caught up to be quite close to the accumulated budget figure to date. There has been no change of personnel. We predict that the budget and actuals for this line item should now reasonably track each other for the remaining milestone periods.
2. Consultants/Sub-contractors
The actual cost of this line item is also higher than budget, again partly due to the inclusion of costs previously missed, but which came to light during the audit review. However, the total actual cost accumulated to date is still high with respect to the budgeted accumulated figure. We expect the cost for this line item to continue at about the same level, but the difference between the budgets and actual should narrow over the remaining milestone periods.
We further expect that the costs for Travel and Living, and for Equipment upgrades will be less than predicted, allowing for some relief for extra costs in items 1 and 2 above.
3. Travel and Living
Travel and Living actual costs continue to be well under budget, and are expected to continue to be modest. The saving in this area will provide a cushion to possible overruns in the personnel cost area. However, it is also expected that these costs will rise during the last Milestone period due to a campaign to visit all or most of the schools for a last person to person meeting.
4. Other
(a) Use of KNOW facilities
An invoice for the use of KNOW facilities for the total period of the tests, has not yet been received from the Open Learning Agency. It is expected that the invoice for the full 2-year period will be received soon, to be included in the Milestone 8 Claim. When received, it will complete this element of cost.
(b) Connection to PLNet, BCNet, Ca*Net3
All connections have been
completed and are continuing to operate satisfactorily. Line connection actuals are close to budget.
Connection to Ca*Net3 was achieved by locating the main Broadcast Collection Server at University of Victoria. The Broadcast Server is where the ordering of websites is done. It is also the site there all the collection of the websites if done. Finally the same server regularly checks for changes on the ordered websites and update are ftp'd to a server and Broadcast Encoder located at Knowledge’s Networks Headend for insertion in the Video Signal.
(c) Special Purpose Equipment
(a) Headend Hardware and Software
All head-end hardware and software has been installed at the headend, and the headend has been operating successfully for months. It is hoped, that during the last Milestone period, a unique AIC Encoder will be tested at the KNOW head-end, that is expected to speed up the data throughput by a factor of several times.
(b) School Hardware and Software
There has been no nw equipment purchased during this calim period, but there has been a claim for an encoder cost that was missed in the first milestone. All equipment will be donated to the schools when the project is completed.
We are now very close to the total budget for all equipment; and while no new equipment is anticipated, any small allowance will be kept for replacements and maintenance as required.
(c) Upgrades
No hardware upgrades have been purchased in this claim period. However, there is space for 4 additional SCSI disk drives and several additional decoder cards inside the Tower Chassis, and up to 512 MB of ECC RAM. A total of 12 additional disk drives can be added externally.
Software upgrades continue to be made as they are released from time-to-time.
At this time, it is not
anticipated that this budget will be used unless a new requirement arises that
we have not anticipated. If unused for
upgrades, this budget could be used for additional personnel work.
In summary, we recommend that the Worksheets and Budgets remain as revised in 01 April 2001, and that further adjustments be made after Milestone 7 to ensure that the actuals and budget will be very close at the end of the Project.
4.0 Decisions required for March 31, 2002 – the end of the Project
The current Contract Agreement calls for all work to be completed by March 31, 2002. However, this date does not co-incide with the end of a school year, and there has been some concern expressed about what will happen to the service for the months of April, May and June.
As far as the head-end is concerned, we expect that, for at least the 3 month period, the Open Learning Agency (via KNOW) will continue to support the transmission of SchoolWeb data-in video as well as host the local server and encoder, and monitor the data in their Master Control.
At the schools, the SchoolWeb severs will continue their high quality of performance, serving up both dynamic and Policy-Based Caching at very impressive rates.
What will be missing, unless further financial support can be found, will be:
1. On-going technical and help-desk support
2. On-line monitoring of server performance and immediate service response
3. Replacement of parts as required on failures
4. On-going software development and upgrades
5. Host server support for ordering, classifying, categorizing …. Web sites
Certainly, Advanced Interactive Canada Inc. and Schoolweb Systems Inc. will be looking for continuing funding assistance to carry on those services. In this regard, we would appreciate any assistance that CANARIE could provide to help us secure such funding help.
5.0 Over the first 3 months of 2002, we will attempt to provide additional training on Policy-Based Caching to each of the schools. In this way we hope to raise the level of use and awareness of this method of caching, to make greater use of the concept, and to provide us with more useful statistics for the final Report.
Respectfully submitted,
Harry K. Davis, CEO
Advanced Interactive Canada Inc.
