Tuesday, November 5, 2013

From Global Knowledge Clickbait to Discourse on Cisco's UC Domination

Bopping around on Twitter Monday, I ran across this tweet by Global Knowledge:

I figured I'd bite, but quickly determined that this whitepaper does NOT tell you how Cisco carved anything in anything.  It's just an overview of the CUCM solution.  This got me thinking it would be fun (Am I weird? Don't answer that.) to write a post that took into account all the things that made Cisco the leader in the Unified Communications space. So here goes my attempt at explaining the story.

Back in the day, you had the PBX.  And the PBX was a MONSTER.  It was complicated to manage, really expensive (both up-front and for recurring maintenance), and inflexible as all-get-out.  Initially, PBX services only augmented services provided by the PSTN, namely dial-tone.  Then came voicemail solutions, and ACD solutions, and ring-downs, and you get the idea..  By the time the mid-90's rolled around computer networks, or Local Area Networks (LANs),  were becoming more mainstream and leading the vendor pack was Cisco.  Cisco was founded in 1984 as a network vendor that sold routers, used for connecting disparate network segments (and sometimes disparate protocols) together, over various transportation mediums.  These networks could be local on the LAN, or remote via circuits provided by the telcos.  In, 1993, Cisco entered the LAN switching market by acquiring Crescendo Communications. Over the next few years, Cisco rounded out the LAN switching porfolio by acquiring Kalpana and Grand Junction Networks.  With this dominance in computer networking, Cisco began to earn a loyal IT following because their solutions JUST WORKED. 

In addition to the Crescendo acquistion, another interesting thing happened in 1993...the creation of the Cisco Certified Internetworking Expert (CCIE) certification.  Thus began Cisco's dominance in another area: mindshare in being THE certification for the networking industry.  Cisco certifications are valued even by Cisco competitors, such as Juniper and Brocade.  Most IT departments view Cisco certifications as a requirement for positions ranging from Help Desk Technicians to Network Engineers.  By creating their certification program early-on, they got the jump on other possible industry giants that might have wanted to create as reputable a certification portfolio.

During the mid-90's, the Internet Protocol (IP) was becoming the de facto standard, and with it more and more applications as well as traditionally isolated solutions were being ported to run over IP-based networks.  One of those solutions was the PBX.  Now in most companies the PBX was owned and operated by the facilities department.  Because the PBX was a huge capital investment, akin to things like real estate and furniture, it made sense for the onus of the PBX to be on facilities.  Keep that in the back of your mind for a bit, we'll come back to it.  The first IP-based PBX was Selsius System's CallManager.  Selsius was formed in 1997, out of another company called Intecom.  Cisco took note of what Selsius was doing with IP-based voice, and decided to make an offer for them.  They successfully beat out Nortel for the bid, and acquired Selsius in November of 1998.  Oh how things would have been different if Nortel had won that bid....  If you want to read more on some history here with Selsius, check out this old post by Mark Nelson, a Cisco employee.

With IP-based solutions, the IT department started to own more and more corporate infrastructure.  Network engineering jobs saw an increase in demand, which required people with Cisco certifications like the CCNA, CCNP, and CCIE.  Therefore, Cisco's network market share increased because these certified folks wanted to work on the same gear they studied for.  See what they did there?

After the acquisition of Selsius, Cisco spent two years getting the product ready for prime-time.  In 2000, Cisco released version 3.0 of Cisco CallManager, which was the first Cisco branded release of the product.  Remember the facilities department and their grasp on the PBX?  Well, CallManager was pitched to IT departments, specifically NETWORK ENGINEERS.  This effectively supplanted the facilities department's long-held PBX vendor relationships.  Well played, Cisco, well played.   This was really a perfect storm.  Cisco was the IT department's friend, because their stuff "just worked".  Ever hear the phrase "Nobody ever got fired for buying x", where x equals a big IT vendor.  Well, because the IT department was being handed more and more solutions, it made sense for them to take the least risky path.  Even with the bugs that plagued those early CallManager releases, Cisco's world class support was there to allow those customers to rest easily knowing that they were in good hands.  This path of least risk is further proven by this brand survey, which puts Cisco at the top of the heap.  The author specifically calls out risk aversion as one of the main reasons for Cisco's leadership.

A strong component of the Cisco strategy has always been to sell and deliver solutions through the use of the Value Added Reseller channel.  The channel acts as an extension to Cisco's sales and engineering staff, without incurring direct expenses associated with the relationships.  Cisco certifications also play heavy into the channel space, because in order to be able to sell certain Cisco products, the Cisco VAR (also referred to as Cisco Partner) must have engineering talent on staff for the products in question.  Cisco Partners often sell other technology solutions as well, including those made by Microsoft, HP, IBM, etc.  This means that oftentimes, the partner sales rep has relationships with accounts that Cisco's own sales team might not have.  Talk about hedging your bets.  This channel delivery model has without a doubt been one of the key components involved with Cisco's dominance in various markets, including Unified Communications.

Oh, there's one more thing that Cisco did to help solidify their position as a market leader. Ever watch an episode of 24 and noticed a Cisco IP Phone? What better way to subtly influence a society full of TV-watchers than to put your products in the middle of TV shows! Genius! This started back in 2005, which was still relatively early for Cisco's UC portfolio.  But it got the image of those phones out there, you know, saving the world and stuff.

Everybody got the recipe for UC success?

Build a brand on a reliable premise, such as the plumbing of your corporate IT infrastructure.  Create a "club" mentality with certifications that aren't that easy to obtain, and that are respected by IT departments.  Encourage people to get said certifications because they will get you network engineering jobs.  Buy a VoIP company.  Stir new and old product offerings until well mixed.  Leverage partners for extending sales and engineering teams. Sell to your certified network engineers in IT departments everywhere. Make tons of money.  Sounds easy, eh?  Thanks for reading!

Monday, November 4, 2013

Some Cisco UC Basics, part three

In the olden days (read: 3 years ago), Cisco's desktop IM/Softphone client combo was called Cisco Unified Personal Communicator (CUPC). The experience was less than stellar.  In addition to CUPC, Cisco had an attempt at a softphone on iOS and Android called Cisco Mobile.  Nowadays, the desktop and mobile device landscape for Cisco entirely centers around a client they call Jabber.  You might remember a few years ago that Cisco bought Jabber, Inc (the folks behind the XMPP protocol).  It's only been in the last couple of years that Cisco has visibly capitalized on this investment, which has basically entailed a major re-branding and consolidation of CUPC and Cisco Mobile.  Jabber works on Windows, Mac, Android, and iOS (no love for Linux, despite a ton of people asking for it, and the irony that all of the UC appliances run on top of Linux).  Do not be deceived, these Jabber clients are in no way the same code base.  A major shortcoming of Jabber to date has been a lack of feature parity between platforms.  That's getting better, but still has a ways to go until the features on Jabber for Mac are the same as Jabber for Android.  Another change in the move to Jabber is that Cisco switched from the presence protocol of choice being SIMPLE to it being XMPP.  That's a little ironic because Jonathan Rosenberg, the creator of SIMPLE actually works for Cisco.

The most exciting thing that Cisco has done in the last few years is purchase Tandberg. Which, as a side note, does give a clue to Cisco's primary innovation mechanism: acquisitions.  Tandberg had really perfected video conferencing through the firewall through the use of the H.460 protocol, which is H.323's solution for NAT traversal.  Basically Tandberg had these two appliances, called Video Communications Server (VCS), one sitting inside the firewall, one outside the firewall or in the DMZ.  The one internal is constantly connecting to the outside box over TCP asking if it has any calls for the internal endpoints.  If it does, then media and signaling ports are opened up by the internal box as requests to the external one, and magically the firewall doesn't have to have a bunch of pin holes added from the outside in (except when the outside VCS is sitting in the DMZ, where you'll need signaling and media ports opened from the outside to the DMZ).  Tandberg adapted this to work with SIP as well, so B2B video "just worked".  You would still have STUN, TURN, and ICE on the outside for endpoint purposes, but NAT traversal at the head-end works through this pairing of appliances inside and outside of the firewall.  Fast forward to now, and this coupling of appliances is going to be a major strategy for Cisco in providing VPN-less connectivity for the Jabber client, but also B2B audio and video calling.  If you're a burgeoning UC admin, I highly recommend getting familiar with the concept of SIP DNS SRV records and SIP URI's, because it's going to become more significant in future UC deployments, regardless of if Telepresence is in use or not.

The Tandberg acquisition happened around the same time that Cisco purchased another company you may or may not have heard of before: WebEx.  WebEx had created an industry all on its own around web conferencing, and with Cisco's acquisition, everyone stood around and scratched their head trying to figure out what Cisco's play was going to be.  So we waited.  And waited.  And then finally something interesting happened...Cisco created an integration option with their traditional conferencing platform, known as MeetingPlace, and WebEx for scheduling and sharing of web content.  MeetingPlace was a monolithic platform that required multiple nodes for audio, video, and web conferencing, as well as separate nodes for web-based scheduling.  MeetingPlace would integrate with CUCM via either H.323 or SIP.  So Cisco's integration between MeetingPlace and WebEx worked OK, but felt like an awkward blind date.  The integration was a little brittle, and the administrative interfaces didn't look anything alike, so it was a bit of a learning curve to juggle the two of them.  A little over a year ago, Cisco announced something that surprised a lot of people: Cisco's WebEx in an on-premise package called Cisco WebEx Meetings Server (CWMS).  This also marked a lack of announcement for another release of MeetingPlace, so the writing was on the wall that Cisco was going to sunset the MeetingPlace product in favor of WebEx.

CWMS is an exciting premise, because it allows for a customer to have a full copy of the WebEx platform hosted on their own servers.  The scalability ranges from 50 concurrent users all the way up to 2,000.  A downside to "on-prem WebEx" is that a lot of compute power is required to run even the basic 50 user configuration.  Also, the solution requires that VMware vCenter is deployed, which can add some complexity to the environment.  The full CWMS solution consists of an Internet Reverse Proxy (IRP) virtual machine, an Admin virtual machine, and for larger implementations (250+ users), a separate Media virtual machine.  For the smallest size installation, a single physical host can handle an instance of CWMS with an IRP and Admin VM.  To make it redundant, you add a second physical host.  As you scale up, more physical hosts are required because the processor and memory requirements grow quite a bit.  At the 800 user mark, the total number of cores required for a NON-REDUNDANT configuration is 80, and the amount of memory required is 116 GB.  At the 2,000 user mark, there's an additional virtual machine type for web conferencing only, and you are required to have 3 media virtual machines.  Total number of cores for a non-redundant configuration is 160, and total amount of RAM is 276 GB.  Ouch.

If you want to connect your video conferencing solution to WebEx, you might want to sit down.  WebEx was a proprietary platform when Cisco bought them.  So proprietary, in fact, that it took Cisco 5 years to announce that they were going to allow integration between their Telepresence portfolio and WebEx.  The integration was named WebEx Enabled Telepresence (WET), and requires a Tandberg Expressway/Control VCS pairing, DNS SRV, the Telepresence Management Suite (TMS) product, a Multipoint Control Unit (MCU), and some endpoints.  The downside to this solution is that you'll likely need to hold an IPO just to afford everything.  To setup a meeting, the conferencing is scheduled either through TMS or WebEx, and when it's time for the meeting all the local participants will connect to the MCU as well as the WebEx "cloud".  Everyone can see everyone else, however to those on-premise folks, the WebEx video will appear as a single feed and vice versa for WebEx participants.  The experience can be a little frustrating if you have a large number of people on WebEx video.  This solution DOES NOT work with the on-premise CWMS.  At least not yet.  

I hope this has been helpful to give you a snapshot of some current Cisco UC basics.  Let me know if there's anything you'd like to read about that has puzzled you, and I'll do my best to break it down.  Thanks for reading!

Friday, November 1, 2013

Ultra-teeny-tiny-book and Choosing Your Own Device

I should preface that I work for Softchoice Corporation, but this blog post is ENTIRELY my own opinion.

Also, one or more of the vendors in this post asked my company to sanitize their version of this blog post, removing all negative commentary.  The following is all the content they didn't want you to see.  Enjoy!

(UPDATE: I'm sending my device back.  It's too under-powered for me, and I'm just not productive with it.)

Ahh the UltraBook.  Small form factor, great good battery life, decent specs.  I've been trying out the Lenovo Helix, the tablet/laptop hybrid that made a splash at CES earlier this year.  My employer, Softchoice, has been trialing a Choose Your Own Device campaign sponsored by Intel and some other vendors, and I was lucky enough to be included in the program.  I chose the Helix because the specs listed in that were better than the rest of the devices.  I'm a power user, so I need at least a Core i7 and 8 GB of RAM.  Storage isn't a big deal, since I keep the majority of my files on Box.net, and can pick and choose what gets synchronized down to my device.  The Helix as it was listed indeed met my spec requirements.

After a small wait, the device finally arrived. And I was like a kid at Christmas time.  (Admittedly, this happens anytime I receive a package from Amazon too.)  I opened the package, turned the device on, and...squinted to see the screen.  The Helix has an 11.6" screen.  And since I've used 10" tablets for a couple of years now, I thought an extra 1.6" would be a bonus.  Nada.  The resolution on the Helix made using the device as a "laptop" near impossible.  Holding it closer to my eyes like I would a true tablet was fine, but when I tried to use the keyboard (you know, all laptoppy and stuff), I looked like a sad T-Rex.  So that's "-1" for Lenovo for pairing a tiny screen, high resolution, and a keyboard without also including a coupon for a visit to an eye doctor prior to shipment.

Then I tried the device's tablet mode.  Undocking the screen from the keyboard was like the starship Enterprise going through a saucer section separation. It just worked, and there were no problems whatsoever.  The device has USB ports as well as a display port on the bottom edge of the tablet, which is normally flush with the keyboard dock.  I definitely love the idea of being able to plug a USB device into my tablet.  So a "+1" to Lenovo for building the tablet-only portion of the Helix to stand alone nicely without the keyboard dock.

At this point, the Helix had been on for about 30 minutes.  I started to notice some heat radiating from the back right, so I placed my hand on that part of the device.  The temperature was uncomfortably hot for a tablet, and I had not really even been stress testing the unit.  Another "-1" to Lenovo.

I decided to see if anything was using my processor, so opened Task Manager.  Lo and behold, I only had 4GB of RAM.  This was confusing to me, because I specifically remembered seeing 8GB of RAM in the spec sheet.  I checked the processor: Core i5.  And lastly, for giggles, the HD was a paltry 100GB instead of the 180GB advertised, though the real disappointment here was on the processor and memory specs.  Giving my company the benefit of the doubt. since this program is a "trial run", I decided to move forward with the usage of the device to augment my desktop.

As an engineer, I can safely say that 4GB of RAM is almost useless in a world with tabbed browsers, Windows 8, Office 2013, and multiple office files nearing (and in some instances exceeding) 10MB in size. Thus began my frustrations with the Lenovo Helix.  I used it for all of 30 minutes on my first day of ownership, primarily because the Helix was just sluggish compared to my primary machine.  There was no way I was going to sacrifice productivity when I could use something else that was snappier.  Ok, so maybe I'll just stick with using it on the couch and when I travel.

This was also my first foray into Windows 8 on an OEM device.  I'm less than impressed with Microsoft's latest OS.  The interface isn't very intuitive and doesn't flow well between applications, especially if you have a mixture of Metro (full-screen) and traditional (windowed) apps. The Windows 8 store is a might....challenged (staying politically correct here).  The majority of the pre-loaded apps wouldn't update, nor could I install anything new from the store for at least a few days.  I think a Windows update sorted the problem out.  The app ecosystem is a bit light as well.  Some of the core apps that I use on my Android tablet were duplicated here, but the functionality was in some instances crippled.

My first road trip with the device was a success, other than the aforementioned issues in using the device as a laptop.  When I arrived back home, using the Helix had made me appreciate my main machine.  I can safely say that given another chance/choice, I would go with something that had a larger screen and better specs, even if I had to put a little skin in the game with my own money.

The idea of allowing users to choose their own device from a pre-set list is a much better approach than allowing carte blanch into your organization (like what happens with BYOD).  In the context of a CYOD model, I would encourage every IT department to work with the lines of business to establish user profiles that equate to classifications of specs.  This way, you're not giving engineers machines that are under-powered and not capable of running VMWare Workstation or opening a 200MB Excel document.  Conversely, it doesn't make any sense to give someone in sales a Core i7 with 8GB of RAM.  Use cases will set the criteria for those device lists.  Also, build your CYO policy to account for what happens when a user chooses a device, and that device falls short of the user's needs.  CYO is another tool in the utility belt of IT to accommodate the changing workforce.  In my opinion, it should always be chosen over BYOD, as it gives control back to IT, and likely will end up being the "best tool for the job" for IT departments everywhere for the next few years.