Mobility and Wireless

Welcome to this week's topic - Mobility and Wireless! Allow us to introduce ourselves, I am Dave and joining us will be Omer, Danny and Tom. Omer and Danny work in Israel while Tom and I work in Northern California. This means that there is never a time when all of us are in the office at the same time so we must rely on technology for us to collaborate.

All of us work in the Intel IT department; Omer and Danny are the leads in the engineering team responsible for deploying wireless technologies inside of Intel. Tom is the lead in the engineering team supporting the client software for wireless, and I represent the Mobility and Wireless technology development group within Intel IT.

In order to keep this blog manageable, we will be using the term Mobility interchangeably with Mobility and Wireless.

What does Mobility mean to Intel as a large enterprise consumer? Mobility is more than a 'nice to have' at Intel. It has allowed us to use the word 'work' as a verb, not a noun - Work is now something we do, not somewhere we go.

Are we walking the walk? - Almost every Intel location has WLAN installed! Also, three quarters of Intel's employees are using Centrino notebooks instead of fixed desktop computers. And we have a campus where not only do we have WLAN installed, but we are changing the usage model by making wireless the primary network access method and using the wired connection as a backup method. - More on that in a subsequent post. Wireless is more than just WLAN, it encompasses all wireless technologies that are globally available - Wi-Fi, WiMAX and a variety of cellular technologies.

In order to mobilize Intel though, we have had to overcome a few challenges; one of them was the security around wireless. Another - usability - this includes the capabilities of the application software. Still another is the challenge of deploying a Wireless LAN to support a large number of consumers (100s to 1,000s) on a site.

So what does my 'mobility kit' look like - A Centrino Mobile Technology powered notebook and a 3G cell phone which receives my corporate email securely. - As a matter of fact, I am working on this post while mobile.

Please keep in mind that we represent the IT department, not the product group, so we can't give information on future capabilities. But we do think we are in a more unique position then most IT depts. - Our users are the people who actually developed the technology we support.

Two of us are located in on the West coast of the USA and two in Israel and we will be responding to your questions and comments on a daily basis.

We are looking forward to your questions.

If for some reason you have no questions for us, we'd be interested in your response to a couple of our own:

• Are you looking to move your internal employees to Wireless?
• Do you feel that Wireless security is robust enough to protect your production information?
• What challenges do you face in implementing Mobility?
• What changes would you like to see in Mobility?

Compatability and OSS

That is great, but what all of us want is open source linux drivers.

What is Intel going to do to help create Open source Linux and BSD drivers for all intel wireless chipsets?

Re:Compatability and OSS

and none of those projects cover what is talked about in the article/story. None of those OSS projects backed by intel support the Cellphone wireless connectivity. Right now you haveto do really advanced hacks to get the cellphone boradband to work under linux (except for non intel based chipsets used with some of the Verizon EVDO data service)

When will Intel start releasing ALL tech and example code for the wireless broadband chipsets so that a native driver can be written? or better yet, drop the bad design of software modems and make them 100% hardware with a fully open and documented interface so that one can be easily written. I should not have to sacrifice any of my processor cycles to do the job that the chipset should be doing.

home user

as a home user, i cant say that i am looking to move my employees to wireless, but i can say that many members of my family use wireless netwoking and that the onboard wireless really helps, it reduces the amount of used ports on any given laptop. the current security system has proven that (at least with wpa) that it is fairly robust. wep has proven that it can be cracked in under 5 minutes or something like that, but wpa appears to be holding fairly well. however: i do not feel that one should rely on the security as a be all to end all setup. any determined hacker can get through any security. so security is better left in the hands of the users than the hands of the software.

Trouble with Intel wireless

I worked for the IT department of a school district. We had a Novell based network with roaming profiles, which made a few things more difficult than average. We also used a unified image, with addon images specific to the hardware.

We bought over a hundred Dell laptops, and paid a premium for the Centrino branded laptops with Intel network cards. For every ten laptops, there was a cart they would sit in and charge, and a dedicated access point that went wherever the cart went.

Getting the Intel Proset software to install as part of our imaging process and to connect before login (to allow wireless authentication,) took a disproportionately long time. It almost made it seem like a waste as opposed to installing it individually on each laptop. In addition, we suffered from connection problems for close to a year. We frequently updated the proset package, did wireless site surveying, experimented with different settings, all with no results.

I still don't know whether it was a driver issue, or a hardware issue with that particular card. Finally, we were forced to "downgrade" to another network card. Once the Intel cards were replaced, the connection problems went away. The drivers which came with the replacement cards were minimal, ugly, but completely effective. They allowed a connection to be established prior to login without any hassle, and worked seamlessly with our imaging process, without any need for extra reboots, automatic logins or batch files.

I will admit, every network is it's own little niche with it's own set of problems. I have spoken to several people who have (and personally have,) used Intel wireless devices with success and ease. I think, though, that Intel's priorities should change in this area.

Utility and reliability should come well ahead of user interface design, and UI niceties which impede on the first two should be strongly reconsidered.

A couple of questions

First off, is Intel specifiaclly waiting for the .n transmission to be ratified, instead like other manufacturers releasing a draft-n chips? (Too bad you didn't have any with the inception of the MacBook - now im stuck with a Atheros chipset for which no (good) free drivers exist atm) Also, what about a integrated cip that does both Wifi and Bluetooth using the same antennae? Also, killthe killswitch! Make the hardware recognize it isnt used anymore, or at least move this to the drivers. Notice no Macs _ever_ had a killswitch for Wifi (or just about anything else), because the OS handles that. Oh, and i wish intel would start enforcing the move away from BIOS somehow. No laptop should even have a BIOS configuration - its not like you need it to reconfigure your new PCI Interrupts or whatever! And it would fix all the broken ACPI shenanigans that some of us experienced (im looking at you Fujitsu-Siemens!).

Mobility?

Laptops today aren't mobile. Nowhere near it. The fact of the matter is if I can't get a full day's work out of a laptop it's not mobile. It's simply portable and usable when you're stuck out somewhere.

Manufacturers need to get out of this rut where "2-4 hours" of battery life is good enough. Look at Dell. The only laptop that gets more than 8 hours is the 12" XPS. 6 hours should be a minimum in this day and age using a base spec (low speed C2D with integrated graphics or low specced discreet graphics) while 8-11 hours should be the norm.

Take the iPod for instance. This one has gone from 5 hours of battery life to 24 hours of battery life. When I went from my 3G to my 5.5G the difference was like night and day. I could operate 3 days without having to worry. Sometimes I've gotten close to a week out of the thing! Laptops need to be more like that.

What large benefits are realisable?

I have two questions.

Question 1: What benefits does business get from wireless, that are worth having?

Considering one of my past workplaces, everyone had large desks in shared offices. There was also a central area with food, newspapers, comfortable chairs etc used by employees at lunchtime. My computer was a desktop with a large CRT. I assume this company was in a fairly similar situation to yours, when you say "Almost every Intel location has WLAN installed! Also, three quarters of Intel's employees are using Centrino notebooks instead of fixed desktop computers."

I'm struggling to think of the benefits of switching from wired desktops to wireless laptops in my employer's situation, for the following reasons:

* It's likely I'd remain tethered to my desk, since I could have a large monitor/full size keyboard.
* I could carry my laptop to the central area with comfortable chairs, but I'm sceptical as to the benefits this would deliver the company - otherwise our offices would have been built to look like the central area.
* If everyone took their laptops to meetings it would create a 'picket fence' with everyone behind a screen, not making eye contact and so on; and I do fine at meetings without a laptop.
* I'm not looking to take my work home with me after hours, thank you very much.
* Laptops are inherently less repairable, tend to offer less performance per $, and may be more vulnerable to theft.

So, that's my question: What is the big benefit of laptops/wireless in an office environment?

Question 2: How do you secure your laptops so Intel IP doesn't get stolen if a laptop gets stolen, while at the same time making data recoverable in the event of a hardware failure?

Thanks!

Michael

Intel's primary WLAN campus project

Hi all,
I want to dedicate this post to our Primary WLAN Campus project (PWC). This project that we do in collaboration with Cisco is a technology study activity that aims to come up with a design allowing complete buildings/campuses to move all users to WLAN as their primary method of network access.

There are many aspects to such a big endeavor and we have published a number of white papers on the details so I want to dedicate this post to the overall approach and challenges rather than the implementation details. There is a link to all our white papers in Intel IT operations home page at http://www.intel.com/IT [intel.com]
First - why do we do this? The simple answer is that with a highly mobile workforce people spend less and less time in their cube and so provisioning network drops and end points (computers, desk phones and so on) to known location provides less and less benefit and the best way is to create a ubiquitous service offering allowing users full access to their communications needs wherever they happen to be. The PWC efforts started with data, moved to voice and we are now working on video and planning off-campus connectivity as a next phase to be able to meet this vision. There are also cost benefits to this direction but I consider that a 'bonus' rather than the main driving force.

So how did we go about making something like this work?

First we need to consider the challenges: to 'unwire' the users we have to deal with any traffic that is used today in the LAN which means data, voice and video. This is a huge challenge because when comparing the WLAN to LAN the former has some clear disadvantages most notably in the fact that it is a shared media over a rather noisy and unshielded environment. This in turn means we need to pay special attention to security (the whole CIA - confidentiality, integrity and availability), to managing the components and the varying service levels in the face of interference and so on. We also have to support very sensitive traffic that requires quality of service (QoS) controls such as voice over IP (VOIP) which is highly impacted by delay, Jitter and packet loss and also incurs high overhead on the WLAN due to small packets.

After understanding the challenges we needed to come up with an architecture and design that will enable dealing with them: Proper product selection, careful deployment of access points, selection of suitable security solutions and development of management tools and practices are all vital parts of making this work.

Finally we 'hit the road' and started deploying our design in a major campus in the US hosting about 6000 employees. We started small with one building and data services only and in the course of about 2 years grew to a complete campus coverage capable of supporting data and voice and we are now dealing with video and looking forward to extending coverage off-campus and providing added services such as location based services (LBS).

This has been a terrific voyage with its full share of revelations, disappointments and moments of exhilaration. As a technical person I could not ask for more.

So what is your take on moving to WLAN as you main connectivity methods? Are you considering it? doing it? love to hear from you. -- Omer

Also in regards to open source

I'd like to know why Intel refuses to allow for the complete documentation sets to be provided to programmers for their wireless hardware, since they do so for their processors and most other hardware. A particular concern with my firm is security, thus we do not run Windows machines at all, the only operating system used on any computers is OpenBSD. From our servers to our desktops, to our PDAs (which, due to security concerns, have to be plugged into the wall rather than wirelessly mobile), all our systems run OpenBSD. Because of the concern for security, we don't allow wireless, but in the near future my boss hopes to implement an IPSec-based wireless security solution, which would limit wireless connections based on various components, including MAC filtering as well. But even there, he is timid - he refuses to permit a security breach on his watch, since prior to his arrival at the company most systems ran Windows and there were repeated breaches of security causing rather big issues (both wireless and landlined), he came in to fix it and cracked down hard. Patches happen the day they are released on all machines, no exceptions for anyone, web traffic during business hours is filtered when going to non-work releated sites and more. To him, only solutions in OpenBSD are worth considering, because to date there have been no issues while using it.

And well, since you asked, I'll answer those.

Are you looking to move your internal employees to Wireless?

No, to date the company I work at not only has a policy of no wireless, they have had us set up jammers for both cellular and wi-fi, because is is such a massive productivity and security issue for the pair of technologies. My boss may be seriously looking to set up limited wireless allowance, but I will only be implementing his design, I'm not doing any looking myself.

Do you feel that Wireless security is robust enough to protect your production information?

No, WPA is no better than WEP before it - to date no wireless technology is secure enough to merit the minor convenience it can provide in a small number of fringe instances.

What challenges do you face in implementing Mobility?

None at all, I agree with the stance of my superiours in this.

What changes would you like to see in Mobility?

Security, reduced power consumption, complete transparency in the functionality of the hardware and all protocols.

Intel 537EP Modem Drivers for Vista

David,

Thanks for being here to help answer questions. I see there is a lot of talk about the newest technology at Intel, however, there is still a group of people still using the good old modem up here in rual upstate New York. Is there going to be any drivers for the old INTEL 537EP chipset? I tried to install an Intel 537Ep chipset modem and it didn't work. Microsoft is nowhere in this either.

Robert

Re:I like Speed!

to put it bluntly, if you would rather have a 3.2Ghz Pentium 4 (even dual core) over a 2.2Ghz Core Duo, you need your head examined, and should not be making technological decisions for your company.

While the 2.2Ghz core duo /is/ a lower clock speed, it is insanely more efficient, and will outperform any P4 that was on the market (disoucinting heavily OC'ed P4s with liquid nitrogen cooling systems).