Currently viewing the tag: "SON"

Ever see the movie Groundhog Day starring the comedian Bill Murray? If you haven’t seen that movie the basic idea is that the main character wakes up every day at the same exact time to the same exact day….

Groundhog-Day-Posters Here’s the deal….

It’s Groundhog Day for LTE!

All of the carriers are doing the same old things, such as playing favorites with vendors instead of truly evaluating what’s good and needed, or performing design, optimization and operations the same way they have done since they were handed the script in 2G times.

It’s unsustainable.

Crazy amounts of wasted time and money are flowing into and around these networks and that is artificially keeping pricing high to the end consumer in a round about way. It’s weird because the operators themselves realize they need to trim up but they are sticking to the same ole same ole. It’s really close to the definition of insanity.

I get frustrated running into the same problems over and over again, as if last year’s lessons don’t apply today, and it doesn’t matter who’s name is on the door, same issues.

 

Let’s stop designing networks for noon users and spending emergency dollars for night users. It’s backwards. Let’s stop locking into a single vendor as if that’s a good idea anymore. It’s not. Dedicate yourselves to IOT testing and force the vendors to step up. The time you are losing waiting for the ‘preferred roadmap to materialize is far more than the time you delay testing a new vendor.

 

PrintRecently Verizon made some waves by announcing their intention to deploy LTE Release 10, specifically the Carrier Aggregation (CA) feature. Naturally I have lots to say on this topic so I thought I would forget what I wanted to say and just hypothesize and prove if true or not. So today is the hypothesis.

Verizon’s over the air (OTA) efficiency is very low with their existing 700MHz deployed channel, and this will not significantly improve without any changes. Therefore, my hypothesis is that when they flip the switch nationwide (depends on device/UE availability of course)  that only people living in deep fades will realize a significant gain… and furthermore, AT&T’s OTA speeds won’t seem that far behind with their 10MHz vs Verizon’s 20MHz. That’s pretty bold but I have my reasons. I will close today with a very typical observation… a clue, from Verizon’s network. Screenshot_2013-12-12-17-44-24 Thanks to my handy WINd tool- I can make these measurements 24×7 with no hands ma…:)

I’ll try to be wrong or right tomorrow!

 

the-secret-to-happiness Update: Serendipitously, Kevin Pritchard at GigOM just published an article on Verizon’s New network. I added some thinking about this below…

There’s disappointingly little to report from the Small Cell Forum meetings in Dallas, TX this week, was hoping for someone to make bold moves. Anyhow, ever notice that most LTE macro network operators are very hesitant to push the envelope as far as performance is concerned. It’s really surprising given that EVERYONE is deploying LTE, devices are more or less ubiquitous and so that really leaves performance and price as the key mobile data service differentiators. 

So that brings me to some recent press release I read about Root Metrics performance comparison of the US operators. Here is a sample of their data:

Average download and upload speeds by carrier 

AT&T: 17.0 Mbps download/7.6 Mbps upload
Verizon: 11.9 Mbps download/5.0 Mbps upload
T-Mobile: 10.9 Mbps download/4.9 Mbps upload
Sprint: 5.7 Mbps download/2.5 Mbps upload
Cricket: 0.6 Mbps download/0.4 Mbps upload

Let’s re-frame this data a little bit. Here is a comparison of Now to Then…

comparison_LTE_may12-June13

So what gives???

Verizon admits it can’t handle LTE demand in major cities (Nov-2013)

AT&T Is Deploying Intucell’s SON Technology as Part of Latest Wireless Network Upgrades (Feb-2012)

SON eh? There are other things they are doing but there are lots of things they are NOT doing either….

comparison_LTE_may12-June13_eff

To be fair, there’s a little uneven playing field here- Verizon has more extensive LTE coverage and more users, therefore you could argue the loading was not identical but on the other hand, with SON on your side, that’s is always true. 

Update!

As I mentioned above, Kevin Pritchard at GigaOM published an article titled “Verizon quietly unleashes its LTE monster, tripling 4G capacity in major cities”  BTW there is a similar article written 15 October 13 By Andrew Martonik in Android Central…Verizon deploying LTE on AWS spectrum in major markets

Here is the summary:

On the third anniversary of its LTE launch, Verizon is delivering a new 4G network. Over the last few months, it’s been quietly deploying the fastest, highest capacity LTE network in the country.

The GigaOM article touts the NEW Verizon network. Hmmm. So let’s look at this 2 ways. Firstly, for existing UEs, The current crop of RF baseband processors in devices like the Qualcomm 9615/Snapdragon 600/800 etc… support up to 3GPP Release 9, therefore most of the devices will only support 1 frequency at a time. I’ll explain why I said that in a bit. This means that old devices will remain on Verizon’s Band 13 (Upper 700MHz) AND only new devices (example: Nexus 5, Samsung S4 USA Version, iPhone 5S/c) will support their new AWS Band. So the capacity crunch won’t immediately feel much better. Next, the best that a user in each band can hope for from the network is 84Mbps peak or somewhere around 11-17Mbps per the tested results above. 

The reason I said 1 frequency at a time is that 3GPP Release 10 introduces a new feature called carrier aggregation. This allows the network and device to bundle multiple spectrums into a logical channel. In Verizon’s case this could be AWS and 700MHz channels sooner and PCS and 800 as time moves forward (and they refarm their existing spectrum.) So the good news is these users will see peaks closer to 150Mbps. I would call that the NEW Verizon network for sure. The bad news is the only device that currently supports that is a Korean market Samsung S4.

Qualcomm has added support for Carrier Aggregation in MDM9225 and MDM9625, so eagerly anticipate a faster Samsung S5 or Apple iPhone 6. So in summary, it’s a bit early to tout that Verizon has a NEW network. Back to my original point, until they get Carrier Aggregation, their spectrum efficiency of 1.19 vs AT&T’s of 1.7 won’t change much!

To the average consumer this means generally, slightly, slower throughput if a Verizon subscriber and Verizon is paying about a 30% premium over the air to provide the bits to you versus AT&T.

So it looks from afar that AT&T’s investment in SON is saving them at least 30% compared to their biggest competitor with identical spectrum. What’s holding everyone back from massive SON implementation again??

Lastly, the GigaOM article mentions that AT&T cannot do the same. Well, they could have if they converted their network to TDD, then they would have an even faster (peaks) network with 20MHz aggregated from existing then they could throw their MediaFlo spectrum (lower 700MHz unpaired block) at it. Instead they will go shopping. They tried to pick up TMO and use that AWS but it looks like the Band 12 guys just cashed in their golden tickets…

Stay thirsty my friends…most-interesting-man-in-the-world

 

Read Full Article →

fig,black,mens,ffffff.u1 My pal Zahid spotted a very good presentation about the Future of Broadband, by Martin Geddes. His post was titled “Bandwidth is not the answer – it’s stationarity” The point of the presentation is that Quality of Experience is still not being respected/monitored. Amen! I have been telling people this for a long time so, it’s good for me to see this elsewhere (I don’t feel like I’m all alone…)  go Zahid, go! Oh yeah, the key thing about this presentation is, keep your eye on what’s important. For example, more bandwidth is NOT more important than a consistent level of service. YES! 

So, in conclusion, this Thanksgiving (USA), I’m thankful that Cisco and Amdocs valued SON, and that there’s someone actually concerned with what is really important. :)

Oh yeah, I’m also thankful that I’m able to work on the next, next thing, that throws everything on it’s ear… more on that later…


Oh, BTW-the cool shirt is @ Redbubble.  

 

 

 

LTE_NorthAmerica_2013 So the show itself was pretty interesting, from the perspective of the sessions. Naturally there are useful and useless sessions. I thought the best sessions were the HetNets. Public safety tracks were pretty good too.  The Network Optimization tracks were antiquated thanks to the technology we were showing at the show (example: see Wilango.com under tools) and some others under development but it was good for those who don’t normally spend a lot of time thinking about this aspect of the universe. As a side note, love to see the look of surprise in people’s eyes when we tell them that they are paying 10X too much while getting inferior optimization accuracy. On balance there was nothing earth shattering but there were a few good exhibits here and there, so time was well spent.. looking forward to small cells, after Thanksgiving, in Dallas, TX.. ( smallcellsamericas.com ) Check out the 2 small cell info graphics at the end BTW…

Here was the full breakdown of tracks:

TRACK 1

Service Innovation

TRACK 2

LTE Roaming/Spectrum Management

TRACK 3

HetNets

TRACK 4

Network Optimisation

 

  • M2M Opportunities
  • Application Delivery
  • Leveraging CEM
  • Enriching customer experience
  • Future evolutions with RCS
  • LTE Voice Options
  • VoLTE in the Device
  • Multi-Mode Devices
  • Devices of the Future

 

 

  • Delivering User Roaming Expectations
  • Making LTE Roaming work
  • Roaming agreements between rural / Tier 1 carriers
  • IPx, international roaming developments
  • International spectrum updates and the impacts on international roaming progress
  • Spectrum Regulation & Harmonisation Updates
  • Building comprehensive HetNets
  • Providing optimum LTE services at large events / arenas
  • Monetising Small Cells
  • SON for Small Cells
  • WiFi and Small Cell Convergence
  • Providing Apps & Services on HetNets
  • SDN / Cloud network innovation
  • ON
  • Policy Control and Charging
  • Leveraging Big Data Analytics
  • EPC developments
  • T&M
  • Network Security
  • Precise Timing & Synchronisation
  • OSS Strategies

And Day 2:

OPENING KEYNOTES DAY TWO – Friday 22nd November

  • Public Safety LTE – an update from FirstNet (TBC)
  • Putting the Customer at the Heart of the Network with CEM
  • CxO First Movers Panel Discussion : LTE Advanced: Timelines, Updates and Possibilities

TRACK 1

Monetising LTE

TRACK 2

Public Safety LTE

TRACK 3

Mobile Data Explosion

TRACK 4

The Future of LTE

 

  • Achieving Maximum ROI
  • Pricing Plans and Data Capping
  • FMC
  • Collective Business Models with the OTT player
  • Network Sharing
  • Wholesale Business Models
  • Driving revenues through network analytics

 

 

  • Business model options and developments
  • Rolling out a public safety network in Canada
  • Spectrum developments
  • The use case – scenario and device evaluation – what is required?
  • Ensuring a consistent QoS
  • The role of FirstNet
 

  • Implementing a convergent strategy for PCC
  • Determining the Potential of Network Functions Virtualisation
  • The SDN debate – How, Why When?
  • IPv6 in Mobile Networks:  Lessons Learned and Strategies Forward
  • Building on the evolution towards ‘smart backhaul’
  • Backhaul for Rural Areas
  • Implementing the right mix of technologies to provide the optimum backhaul network
  • Securing the backhaul connection
  • Impact of Non-Contiguous Hotspot Deployment Of LTE On Performance And Throughput And Its Financial Impact
  • EPC evolution – the role of EPC in handling Mobile Data Traffic
  • Service Provisioning and management via PCC
  • Antenna Developments
  • DAS
  • MIMO
  • Active Antennas
  • TD Ecosystem and worldwide updates
  • FD and TD Convergence
  • Carrier Aggregation
  • LTE-A updates
  • Device/chipset LTE-A requirements
  • WiMAX migration

Deploying Metro Cells means Thinking Differently.

Download (PDF, 806KB)

 

Four Step Approach for Deploying Metro Cells.

Download (PDF, 300KB)

don_t_cry_Stunning_photographs_of_animals_in_the_womb-s390x320-80891-475 Have been hopping around the various LTE and SON related standards bodies and thought I would take a shot at organizing all of the links into a central location so that I don’t have to save them in my browser anymore.  The bookmarks are so out of control I’m embarrassed so I’m outsourcing. Anyway, I don’t recall a similar list anywhere so that was another justification.

I will continue to refine. Got some ideas about how to make more usable but that will have to wait for a rainy day…

See the menu above labeled Standards Bodies to see what I mean.

 

 So when I saw the article in the MIT Technology Review yesterday, I knew there would be some hype around it. Fast forward about 24 hours and the hysteria machine has really started up, see for yourself in Engadget and Fierce Wireless posts.

Their hysteria is that LTE networks are easy to jam, using easily procured equipment, the number thrown out was $650. See this quote: 

According to the research group’s director, Jeff Reed, a single malicious operative with a hot briefcase and a bit of know-how could take down “miles of LTE signals.” If the attacker splashed out on an amplifier, they could cut off reception for thousands of people across a whole city or region.

Addendum: This paper was created as a submission to NTIA regarding Public Safety LTE. It is here.

Well that’s easy to take out of context. Any electromagnetic transmission is easy to block/jam when you think about it. So this type of hype brings out a huge irritation with me, and that’s the blogosphere’s copy/paste system to fill their site and help generate page views. If a LASER is Light Amplified by Stimulated Emissons of Radiation, then HypASER blasts are due to Hype Amplification by Stimulated Emission of Ridiculousness

Let’s all get wound up about LTE jamming because hundreds of bloggers will copy and paste the MIT article in various ways to generate some hype for you.

OK, I’ll stop nagging…but the blogosphere is not helpful here.

I’ll explain my logic.

 

 

Applicability

It’s not far fetched to imagine hackers or terrorists or criminals whatever actually doing this. The parts required are mostly off the shelf and the knowledge is easily obtainable. This same approach as described for LTE works the same way with GSM, although CDMA/WCDMA is a bit more resistant but not immune. Creating noise in the RF domain, particularly in the channel of interest happens all the time. For example, an anti N order passive modulation (PIM) war, caused by shoddy work, bad cables, antennae or RF equipment, rusty bolts etc… is being fought now because operators realize the generated noise reduces throughput and thus reducing data capacity and therefore limiting revenue. 

For LTE networks though, the laws of Physics still prevail (in our universe) and a bad person with a jammer will be likely using low power, or having low effective gain (hard to carry around a 9′ antenna all the time), if they are low to the ground (where maximum effect could be achieved.) Again not impossible at low power/gain so you could say the sphere of influence is going to be very narrow if they target the eNB TX band. If they target the eNB RX band they may have more success but it’s effectiveness is wholly dependent on the location of the UE’s trying to communicate to the eNB. THIS IS NOT THOUSANDS UNLESS IN A WELL DEFINED/CONTROLLED ENVIRONMENT/VENUE LIKE STADIUM POSSIBLY AND IS VERY LIMITED IN SCOPE.

Worst Case

Let’s go further and say that the perpetrators have now worked out how to maximize their gain to compete with the nearly 1KW ERP from the base stations. Got to find a favorable (high) location and have lots of gain, so huge antenna or high power or both. What spectrum are they broadcasting in? 

700MHz rogue transmitters may affect larger areas due to propagation characteristics than say 2100MHz. Either way, there is redundancy in most of the mobile world as networks are generally overlaid on a technology basis, so a failed 4G connection moves back to 3G. 

Countermeasures

So I thought it would be fun to review the many existing countermeasures that could be useful in defeating the perps. Firstly there is physical redundancy. Multiple networks, multiple LTE carriers, multiple sites more MIMO (antennae.) More spectrum to cover increases the perp’s setup complexity. They would need to deny 3G networks too. In most cases mobiles could search and find another network to serve them. More sites include wifi and small cells. Small cells alone could be a very very effective countermeasure. They don’t have to be at the same channel bandwidths, MIMO ranks (ex 4×2, 4×4, 8×8 etc…) or could/should operate in different channels or even utilize TDD modes instead of FDD modes or vice versa. This in an of itself would be very difficult to overcome.

It should also be noted a good defense would be detection. Sudden noise rises are reported in the link prior to all out failure. Beyond that it is wise for operators to have monitoring equipment placed in the link to guard against interference anyway. These external monitors help reduce site visits and so on for common unintentional interference could be the canary in the coal mine for intentional interferers. SON could also help. SON controllers would detect changes in noise and traffic levels, if a suitable outage threshold can be defined, then once the threshold is met, SON could automagically change tilts/increase gains/power in neighboring sectors or sites to help mitigate for the subscribers. 

Not out of the question but a little more resource intensive would be doing things like manual intervention. Examples include turning the cells in affected area off, alternative bandwidths/growing multiple channels or switching modes to TDD mode so as to be able to manually locate the noise sources. A more passive but effective countermeasure would be to implement LTE Roaming such that mobiles always have an alternative.

Let’s not forget that Release 10+ specifications (LTE Advanced) include a feature called Carrier Aggregation that allows operators to operate a virtual large channel over multiple smaller ones. This by very definition is more robust to interference than less bandwidth. Your milage will vary of course but it’s helpful.

Interference cancellation techniques are going to become widespread on UE and eNB to dramatically improve performance and this approach could help a lot.

If the perps are capable of ultra wide band, ultra high gain interference then they are probably more like nation-states and you have a much bigger problem on your hands than just the wireless communication interruption…although those small cells are probably still carrying traffic close by…

Thinking through this for Public Safety, heck this story could be created by a large operator trying to prevent Public Safety from operating their own LTE networks but I digress…the standards could be improved to allow for improved control channel redundancy/resiliency beyond a doubt. 

I guess I refute the numbers but not the principles of the original article. It’s going to take a lot more than $650 to effectively take out thousands of LTE users. LTE networks are probably more susceptible to IP hacks than RF hacks. However the blogospheric focus is on the (hot) air portion. Hopefully the hysteria will die down soon. Ugh…

Reverb  Networks just issued a press release (see below) that they were awarded US patent 8,229,363 for detecting interference via SON. The way I read it was the UE submits everything seen during scanning and the sites already in the neighbor list are removed and therefore the reminder are interferers. The overview in the patent text is:

An embodiment of the invention is directed to method for detecting interference in cell sites of a wireless communications network, … receiving signal code power measurements from the mobile devices in the cell site; and determining sources of signals being transmitted to the mobile devices in the cell site based on the signal code power measurements. In addition, the method includes generating a list of sources providing signals to the mobile devices in the cell site and then removing sources included in an active set for each mobile device thereby creating a list of interfering sources. The interfering sources are, for example, sources not assigned to provide wireless communication services to a mobile device in the cell site, but whose source signal is still being received by the mobile device.

Let’s ignore Automatic Neighbor Relations (ANR) or even Neighbor List Tuning (NLT) seen in CDMA networks for the moment. Patent 8,229,363 is going to be helpful to Reverb if they can keep it. This technique seems like it would be beneficial for all, hope they license it FRAND [Fair and Non Discriminatory] style. 

This won’t win me friends but…these guys deserve a gold star for getting this patent. I have looked at probably over a hundred LTE patents for curiosity’s sake and the industry is very lucky I am not consulted by the USPTO before granting patents. On the other hand, I haven’t looked at other similar patents, maybe they did this defensively?

Patent is here.

Reverb PR:

Reverb Networks Awarded Patent for SON-based Interference Detection

Thursday, September 6, 2012

Advanced Techniques Used in Automated SON Solutions

Sterling, Virginia – Reverb Networks, a leading developer of intelligent Self-Optimizing Network solutions designed to provide mobile network operators with improved operational and spectral efficiencies, announced today that it has received a patent award from the US Patent and Trademark Office for automated interference detection techniques using measurements from mobile devices, a key component of SON-based network optimization for UMTS and LTE wireless networks.

Reverb’s interference detection patent, USPTO number 8,229,363, follows the Minimization of Drive Test (MDT) principle identified in emerging 3GPP technical specifications.  Interference sources are typically identified in mobile networks by using test mobiles and receivers in an orchestrated drive test setting, a process that is both time consuming and labor intensive.  Motivated by the MDT principles of automating data collection from OSS and standard device-based measurements, Reverb’s intellectual property outlines a method of identifying and ranking interference sources.  These techniques are incorporated in the Interference Reduction feature of Reverb’s InteliSON®, a fully automated, closed-loop SON solution.

“The interference detection patent award bolsters Reverb’s IPR portfolio in SON technology,” said Magnus Friberg, CEO. “This award, combined with our other issued and pending patents, demonstrates Reverb’s unique technology for automating network optimization with our InteliSON platform.  We will continue to drive innovation in SON technology as we further deploy our leading edge solutions in 3G and 4G networks worldwide.”

 

 This is not a huge deal but I was looking at DSLReports blog post/story titled “AT&T LTE Network Slower in Markets Like Chicago” and I saw a couple of incorrect facts that although I don’t normally feel the need to correct all inconsistencies, felt like I’d pick on this one a little. I have thrown up the 700MHz spectrum chart for your viewing pleasure.

Firstly our buddies at ATT are deploying initially on Band 17, so they will always be limited to 10MHz channels. If they were to own all 3 bands in a single market, which would be rare, they still would only have 18MHz at their fingertips so that would only be good for an almost never used 15MHz channel or 10MHz channel + 5MHz channel.  

Therefore, most of the time, VZ and ATT are pitting 10MHz of LTE spectrum against each other so other factors are determining the throughput.

As I mentioned before, check the following first:

  • Interference environment
  • Backhaul quality and bandwidth
  • Antenna configuration (antenna quality too)
  • UE performance
  • Loading
Remember this chart from before? According to testing from RootMetrics, I don’t believe ATT is leading in spectral efficiency significantly so all things being equal, likely Chicago needs more down tilt to improve throughput….Oh yeah, and although they are rolling out Intucell’s great SON products, they haven’t deployed RET (Ok neither has VZ) on their towers so they will need to optimize the expensive way.
MY $0.02. Have a great day.

I am working with a regional operator’s LTE network design, and I was comparing our design to some of the publicly released information available about the competitors and, well I was a little surprised at the early results. I think I’ve boiled it down to 2 charts.

Firstly, as a matter of review, ATT and Verizon have very similar resources currently dedicated to the deployments. ATT is deploying 2×2 MIMO using FDD on lower 700MHz, Band 17 (which is only Channel B + C of Band 12) which facilitates 10MHz channels with their existing sites without RET antennae widely deployed yet. VZW has deployed/is deploying essentially the same thing in Band 13 (upper 700MHz.) So knowing what we know about 2×2 MIMO we can baseline where they are.

The first chart is a standard SNR vs throughput chart. As you will see most of the results reported are achievable with only 5dB SNRs.

Here is the throughput comparison.

 

The biggest takeaway for me on the forward link, neither is really achieving an expected baseline level yet. Not even close. They are about 26-28% off the median so there’s some red flags to note in their approach. Is it all due to the early rollout phased spotty coverage? No, many of the measurements were made at airports and core areas for VZW (maybe why it’s slightly higher) and for ATT the have limited markets rolled out so the tests are only in the preferred areas anyway.

After looking at the throughput, it’s only fair to derive their efficiency in terms of bps/Hz. Ready?

It’s directly related to the throughput, but it’s still eye popping to see as efficiency. Now granted, I didn’t throw in the targeted because that only a reference number that’s synthesized with McDonald’s patties in a lab and therefore not that relevant.

There’s a small caveat to throw out and that’s the device performance. Surely the device performance will improve over time.

My takeaway here is all those body shops importing engineers from the Middle East, Asia and so on need to speed dial their Verizon and ATT lunch buddies and work hard to get some of the ATT and Verizon optimization project dough. OK seriously, ATT and Verizon seem to have lots of bottlenecks in their LTE networks that are needing optimization. Question is, is network equipment maturity and device maturity all there is to improve here or are there are things to address?

Oh, and just by looking at this data it’s easy to see that these networks are clearly not under near realtime SON control either. And what will these numbers look like at higher frequencies, like for Sprint and Clear? Hmmm…

Nexius is doing a good job marketing their body shop chops with a new LTE info graphic …but they forgot to mention the things above though, oh well….

A Nexius Infographic - Seven 4G LTE facts that might surprise you

VoLTE

VoLTE

Voice over LTE (VoLTE) is the next big thing. In fact, in 24 months, virtually all LTE enabled smart phones will support it. Curious?
Small Cells

Small Cells

Small Cells, previously known as 'femto' or 'pico' cells are possibly a savior to network operators. They offer capacity and coverage to the end user and are inexpensive for the network operator. Why aren't they everywhere?
Public Safety

Public Safety

LTE is and ideal technology for Public Safety use. See Why.