Skip to Content

ITPG Meeting Notes: 2011-02-17

Printer-friendly versionPrinter-friendly version

Attendees (27):

John Ajao, Arlene Allen, Kip Bates, Polly Bustillos, Ted Cabeen, Michael Colee, Mark Dotson, Ann Dundon, Kirk Grier, Richard Kip, Bill Koseluk, Tom Lawton, Dan Lloyd, Jennifer Mehl, Elise Meyer, Steve Miley, Bruce Miller, Alan Moses, Tom Putnam, Fuzzy Rogers, Andy Satomi, Glenn Schiferl, Kevin Schmidt, Jason Simpson, Jamie Sonsini, Heidi Straub, Jim Woods

Round Table Member Discussion

It has been a while since we had a general member discussion.

A few topics and questions to consider:

  • The Symantec anti-virus contract is drawing to a close. While there has been a lively discussion of antivirus tools on the CSF mailing list, a focused effort to find a solution (or solutions) for the campus has not been attempted. Should we improve this effort, or let everyone go off in their own direction? 
  • How or are you using Cloud Services?
  • What's your favorite new tool or application for your department or unit?
  • What's a short list of topics ITPG should consider for the upcoming year?

We started with discussion of the current situation with respect to anti-virus software:

  • There is a UC system-wide contract for Sophos
  • The UCSB annual Symantec contract will be expiring this Spring
  • Some may be using Microsoft Forefront via MCCA, and others may be using Microsoft Security in Windows 7
  • Others may be using Eset

Are we content that groups are going to go their own way, or is there enough interest to form a group to try and get better pricing?

Then we did a round table discussion:

IA (Heidi Straub):

IA is finishing a VM conversion of their development/test servers from 5 physical servers to 1 Hyper-V base server.  IA had been testing VM with an Oracle database server and IIS web server in an VM environment for the last year (plus). The testing proved positive and IA is now just finishing migrating their full development/test environment, in conjunction with IS&C, into a new base server. The servers are not clustered. (A question was asked whether anyone was clustering via Hyper-V.)

IC (Bill Koseluk):

Every year we go through the Symantec renewal process, the cost is always higher.  You often have to talk with multiple sales people before you get the final price.  Sophos is the most prevalent solution system-wide.   

Chemistry (Jennifer Mehl):

She's looking for solutions for presenting calendaring information as event information in a digital sign.  Has anyone thought about consolidating calendaring information for events? 

Student Affairs hosts a campus calendar of events that is open for anyone to post to, and the Office of Research hosts a calendar of seminars, lectures, symposia, and conferences that is managed by hand.  There are a bunch of people doing this around the country.  Be aware of the trade off of spending 30 hours to set up an elegant system vs. editing requests as needed.  Search the Educause CIO mailing list archives for the recent discussion that included both various packages and how to organize the content. (NB: Search for the subject Digital signage.)  Does Berkeley have unified calendaring?

The Operational Effectiveness Initiative has a working group for Conferences, Event Planning & Ticketing (NB: Katya Armistead and Miki Swick are the co-chairs.)  A related issue that they mentioned at a recent meeting, was that there is a Safety aspect to knowing what events are going on in what rooms on campus. 

There are two issues: 1) How to maintain the events calendar, and 2) how to filter it for different customers.  What signage package are people using?  Packages mentioned included
fitpc2 via redpost ubuntu signage system, and android operating little displays.  Check out Housing's system (Julie Levangie is the contact for Housing's Digiknow system).  Communications Services wondered if there was interest in a campus-wide  signage system.  How many departments have signage?  CNSI, some groups within the College of Engineering.  This is also a political issue.

Physics (Glenn Schiferl): They're treading water.

OIST (Jamie Sonsini):

They are looking at the cloud as a new model, i.e., something being done by someone else.  Due to a potential retirement in a year, they have decided to get out of the solaris business.  So they are migrating their solaris services (listserver, proxy server, webhosting for 35 departments) over to a linux environment managed out of the systems group, which is hosted on something, but they don't care.  So they want to train staff to manage services at the top layer.  The target is to complete the migration by end of June 2011.  They are starting to test the Library services now, next will be Alumni services and the listservers.  Mainframe pdf printing is being done via Xythos in a linux environment.  They would like to not have root access, so they are building a new support model.

Communications Services (Bruce Miller):

When Communications Services was annually doing $1M in purchasing, they developed an internal purchasing system that uses infopath forms, sql server, notifications, and work flows.  What will the UCLA FIS bring to the department level?  The UCLA Bruin buy can't do the workflow that the MSI purchasing system can do.  The CS system can do many forms including facilities requistions. UCLA uses post approval (PANS).  This is like the difference between the flex card vs. procurement card.  One justification for the de-emphasis of the procurement card is to use online catalog.  Using the UCLA systems generates incentives.

OIST (Arlene Allen):

UCSB has been doing Software as a Service (SaaS) for awhile, e.g., OACIS.  They are doing feasibility study of whether a private cloud SaaS is workable at UCSB.

Oracle's purchase of Sun has become a new UC-wide subject.  There is an effort being led by Berkeley to put together a UC-wide relationship with Oracle.  Oracle owns PeopleSoft, the former SUN IdM, Solaris hardware & software, Brio (was Hyperion), Oracle calendar, advance.  Based on a purchasing analysis, OIST spends the most with Oracle.  Advancement is also a big customer.  Of these categories (hardware/software/other) What do people think is representative of their views?  Steve Miley bought 2-Sun NASs that are supported through 2012.  He had a bad experience trying to find out about support. Advancement uses vendor representatives for their support.  David Alix has been successful in getting support.  There has been a change in culture when Oracle took over Sun.  We need to be treated like a smaller group, e.g., the current web access for support allows one group to see another group's hardware.  Oracle needs to understand our decentralized environment.  David Alix has passed off his contact information to Shea.  These issues are similar to dealing with IBM & Cisco, i.e., the vendor wants to deal with the university as one a single organization.  VM-ware is the same, but as we aggregated we got better discounts, versus divide and conquer.

IC (Steve Miley):

They are using the following Software as a Service tools: 37 signals, campfire, chatroom service, team rooms, & one on one rooms, and base camp for project management.
If an SaaS product is successful they are always improving and getting better, plus 3rd parties are developing phone apps for it.  They are using zendesk for a help ticket system and the iPad app provides a better experience than a webpage.  They are investigating Joyent for cloud computing, where one can specify number of processors. They are trying to learn how this service handles notifications, documentation improvements, and the level of reliability or support.

Earth Research Institute (Michael Colee):

They aren't doing much cloud stuff.  They looked at external cloud services to try and outsource cpu crunching, but transferring terabytes of data kills the pricing.

Bren (Jason Simpson):
Their internal cloud is moving from vmware to hyper-v.

tools:

windows patch management using ms system center essentials http://www.microsoft.com/systemcenter/en/us/essentials.aspx
also console for ms hyper-v, inventory, server management

using spiceworks to inventory, monitor, document systems information and asset reporting.
http://www.spiceworks.com/
has many other features that i am not currently leveraging. (This allows me to find and investigate the systems that failed on patching.)

ninite all-in-one application installation for common free apps
http://ninite.com/
pro version does extra things like silent install, remove toolbars, http://ninite.com/pro
offline installation etc.
command line ref:
http://ninite.com/help/features/switches.html (It can also remove desktop icons.)

I use psexec (part of the syssinternals suite) to do remote deployments to all the machines we manage
http://technet.microsoft.com/en-us/sysinternals/default
http://technet.microsoft.com/en-us/sysinternals/bb897553

If anyone wants more info on how we are using these tools I am happy to talk to them

OIST (Elise Meyer):

I've been using it.ucsb.edu for collaboration on projects between operational groups and committees.  It's strong point is sharing a document between the two groups once the editing has been completed.

From a personal standpoint, I've been doing more in the Google Cloud and I'm wondering how much self-backups should I be doing to protect myself from self-inflicted problems, e.g., deleting a weeks worth of data.  Some commented that paid versions of the tools come with backups.  Also, I've installed a VZW network extender that provides in-home cellular coverage that is then transported via your ISP connection to the carrier.  It works well, but there is no way to block other VZW subscribers from using it if they get within 15 feet of it.  (You can prioritize numbers, but you can't white list.)  Someone commented that with the AT&T version you can white list particular numbers. 

MSI (Jim Woods):

Their financial/purchasing system is Coopa.  For mailing and news lists they use a free service (but that means that a 3rd party holds mailing list).  They use this for newsletter email blasts for marketing and outreach.  Some noted that Alumni has a database of all alumni email addresses and they can provide departments with extracts based on different criteria, e.g., Northern California alumni.  They also update subscription date.  Talk to Polly.  They provide the service to other departments.  They also purge the list of bad addresses.  They use a tool that helps with research proposals that ties into grants.gov. A tool called Cayuse will submit a research proposal for you.  Contracts & Grant folks did a webinar.  Both UCLA & Cal Tech have switched to it.  It was built by researchers who hated grants.gov  Contact Bonnie or Tim for more information.

Communications Services (Ann Dundon):

topics to consider:

- IT staff training: Training budgets will shrink at the same time that people are taking on more tasks.  There are two types of needs: critical job specific skills training and generalized/core training that we all need to know, such as Net Citizen, IT policies, and Information Security best practices.  The WSG has set a great example for doing that, also Steve Miley's brown bag seminars.  But there needs to be a group to keep it going.  Can OE sponsor this?  Another example is the County of Santa Barbara doing technology days, which are one hour seminars on security.  Can ITPG put together the subject matter? 

In the past, Alx Sanchez (IC) was able to provide "poor man's" training on demand.  He would learn subjects in order to be able to help instructors.  IC is focused with helping students.  This training series got cut due to budget cuts.

The WSG brought Gabe from UCSD to provide web security training for the cost of his travel expenses.  So the WSG is not doing all the work to provide content.  His session raised the level of knowledge on web accessibility.

If we know what the training needs are and/or what people are interested in learning, then we can look for the required resources.  The WSG uses an exit survey to get their next topics. 

Steve commented that when he put together the IT Summer Series, no one wanted to volunteer to talk about their expertise.  People prefer to just show up and talk about what they know, more like BOFs.  Is this an Operating Effectiveness topic?  Yes, we should be doing it.  Supervisors should be aware, so that people can spend a day preparing for giving a class.
Maybe once a month we could rent a venue for 4 hours and talk.  Drupal group meets and works that way.  The ITPG Comm subcommittee was tasked to address this issue.

LSIT (Alan Moses):

They aren't doing any Cloud computing.  They are looking at SaaS (and SaaS roadkill), so they are testing export capabilities.  A service offering that they are looking at are: virtual computer lab, where a reservable image of software can be promoted to computer lab.  They thought that Statspack might be a good test case, but flash-based video is a better test case.  They are using kvm and a scheduler from UNC that is now in Apache project, and UC doesn't like the Apache license.

One side effect of Collaborate - they are developing good cooperation between the various stakeholders (EVC, Registrar, FM, ID) for their classroom upgrades, and they are
expanding wireless into more instructional areas.

OIT (Kevin Schmidt):

The only cloud service they are using is virtualization within kvm.

ECI (Richard Kip):

They aren't doing much external cloud computing.  They are doing a lot of internal cloud computing.  Rich and ?X? have moved their Eucalyptus cluster to K-mart.  The researchers run their own clouds.  They are looking at a centralized cloud using Eucalyptus.  It's Openstack vs. Eucalyptus.  Openstack is the golden child for cloud computing - it uses the same components and technology.   They wonder if the whole thing will be thrown out in 5 years.  Paul Weakliem wanted everyone to know that UC grid has a one-day conference
session coming up at UCLA.  A UCSB faculty member is presenting on their version of Google App engine App scale.  Go to ucgrid.org and look for cloud summit.  They are looking to  support researchers with an on-premise cloud.  Researchers have software requests (from ancient software), so it is desirable to give researchers their own vm space.  His favorite new thing: GauchoSpace, COE is moving to it quickly. 

IA (Polly Bustillos): 

They are in the clouds with Heidi & Jamie.

Library (John Ajao):

  • In-house cloud-Vmware clusters
  • External cloud- Working on a few NSF research projects developing formats for long term preservation and archive of geospatial data (NGDA)
  • Partner with Library of Congress (LOC), California Digital Library (CDL), Vanderbilt on  Research and Education data depot network (REDDNet), on developing logistical networks and housing of it's nodes. Dataone project, archiving nodes.
  • collaborating with CDL on developing micro services, i.e. ingest tools, metadata tools, repositories for digital library and long-term preservation.
  • antivirus moving to sophos, and windows 7 deployment w/public systems in the library.

SIS&T (Tom Lawton):

They are too embedded in Symantec to change this year.  They plan to deploy Windows 7 this summer.  They are doing nothing with an outsidel cloud, but inside they are doing virtual machines.  They have a 100 vm's that can spin up/down based on daily need.  They are snapshotting things, and it's very dynamic.  They are using vmware, and they have an investment in virtual setup web page that can be generated.  They are using Zen to push apps based on who you log in as.  They are trying to get off Netware before the end of the year.  Different people and projects have different tools, and try to standardize tools and use the tool ACE analyst series by OPnet, it was previously called network physics to analyze network traffic.  Now they are analyzing code in the Windows environment.

Academic Personnel (Andy Satomi):

More and more people are using dropbox.  They are using desktop skype via skype & webcams to communicate with UCOP.

Math (Fuzzy Rogers):

His topic to consider is a  BEG white paper on switching to IPv6 - there will need to be things written up about IPv6 for implementation
requirements.  There has already been some discussion on this topic at SEC-WG.

IC labs (Mark Dotson):

cloud stuff - They have been doing physical to virtual migration type stuff for 8 years: the print stations were replaced by panologic, one student open access facility (Phelps 1513) is now virtual desktops, and the athletics ICA study lab is also virtual desktops.  They are using the Advantage virtual remote type access infrastructure, with a USB key that gives you the exact VM you were using, and you can also access same VM from home.  They can do provision on demand.  Currently doing a pilot of a remote virtual lab setup for a Political Science class which needs access to SPSS.

tools - They use LogicMonitor (Steve Francis's company) SaaS install agent on linux or windows that has network access to devices that you want to manage.  It tracks it all for you - switches, windows servers, linux servers, esx credentials and esx stuff.   They are patching with Shavlik; patches all those free apps like firefox, skype, Java JRE, etc. A negative is that as the company grows they include more stuff (like spyware prevention, new tools, etc.) which increases yearly maintenance cost.  They are looking at application virtualization offerings like VMWare Thinapp.  You can run every version of software side-by-side and configure as wanted.  It allows ie6 and support problems they have tons of apps and push out tons of data, to build it and test and then put into start menu shortcut.

Life Sci (Ted Cabeen):

  • They are changing to Eset, so they have started their converion from Symantec. (They excluded sophos.)  They are willing to go with others to get best price on it.
  • Over the last year they have gathered data on the ownership of machines, and have learned that there are lots of personal machines from grad students and faculty.  So they are having people migrate to Microsoft Essentials if it is a personal machine, and buying Eset for university owned machines.
  • They aren't using the external cloud much, but internally they are doing vm, but they don't have auto provisioning.
  • tool: cloud based backup for $6/month will keep infinite number of copies of files for 5 years.
  • They are rewriting host management system.

ID (Dan Lloyd):

  • They are using Eset.
  • They use Matterhorn for streaming solutions.
  • They are getting classroom lecture capture going, and they did the WSG podcast.

OIST (Kirk Grier):

They are aiding and abetting the FIS & SIS projects, and both groups have need for testing.  They are looking at how to leverage their expertise into new services as administrative applications go away, e.g., vmware.  He's learned how you can break out your instance from campus listing from your account your product and contracts.

His suggested topic: What should a central group do as business applications go away, how to reinvent ourselves.  The anticipated FIS & SIS targets are June 2012.

Informational Items

CIO Report - Tom Putnam

There is a mad dash to change everything at once on 7/1/2012.  The big 3 are: SIS (which has their kick off meeting today).  They signed their contract, and are starting converting code from adabase natural, c sharp, sql server, and ?  FIS: where they are still understanding interactions.  The systemwide single PPS and HRIS system, which is rapid acquisition process.  The rfp was issued this week for replacement systems.  The results are due mid-March.  They plan to have a vendor chosen by mid-April.  With the contract signed in May (dis-may!)  The roll out will take 3-4 years.  We don't know what the UCLA FIS is doing with respect to this PPS project.  The places that you pull data from will change.  Someone asked whether the datawarehouse was going to be stable during that time?  Answer: The working group is looking at that.  UCLA has a datawarehouse and an operational datastore, and they are trying to move it all into a datawarehouse in Cognos.  Our datawarehouse is more of a datastore.  There are discussions of keeping our detail in the datawarehouse, which isn't on the mainframe.  A lot of us have things that we do today via the datawarehouse, and in the future we'll be using a different set of tools.  There will be a huge amount of training.  Some of the differences that we'll need to adjust to: we have 7 layers of management in our table, and they have 4.  It's all moving for once, but it is moving.