Wednesday, 30 January 2008

Comparison: HSPA+ vs LTE

Advantages of LTE over HSPA+
  • Flexible Spectrum usage possible with LTE: LTE will be the same whether the bandwidth available is 5MHz or 20MHz. Ofcourse the data rate will increase when the BW is increased. With HSPA+ only 5MHz bandwidhts possible. Similalrly with HSPA+ only FDD mode of operation is possible whereas with LTE FDD or TDD mode is possible.
  • Spectrum Effeciency: Better spectrum effeciency, by a factor of 2 atleast over HSPA+
  • Simpler Architecture: LTE has a much simpler and relatively flat architecture compared to the legacy UMTS network in HSPA+
  • Higher Data Rates: LTE gives DL data rates of 144Mbps and UL of 57Mbps. HSPA+ gives 42Mbps in DL and 11Mbps in UL
  • Ultra Low Latency: 10ms instead of 50ms for HSPA+
  • Short TTI: 0.5ms instead of 2ms for HSPA+

Advantages of HSPA+ over LTE
  • Will be ready much before LTE: HSPA+ technology should be available in Q1 2009 whereas the earliest with LTE would be sometime in 2010.
  • Much less investment in infrastructure: Since HSPA+ is evolution of HSPA which is already being deployed, it would be easier and less costly to upgrade. With LTE since its based on OFDM a lot of new components will be required. Also in case of LTE the number of components are reduced but since they work in a different way, new components will be required.
Did i miss something?

Friday, 25 January 2008

Nokia Mobiles Dominate world

Nokia, the world's largest mobile phone maker, sold four in every 10 phones globally for the first time at the end of last year, boosting full-year operating profits by 46 per cent to €7.9bn ($11.6bn).

Strategy Analytics, a research company, said yesterday that Nokia's 40 per cent fourth quarter market share was larger than the combined 35.6 per cent of its three largest rivals, Motorola, Samsung and Sony Ericsson.

Motorola, the US's largest mobile phone maker, last year lost its position as the world's second largest to Samsung and this week reported an 84 per cent drop in fourth quarter profit to $100m and forecast a loss in the current quarter. But Samsung is gaining on its Finnish rival, saying last week that fourth quarter profit from its telecommunications business, which includes mobile phones, rose 67 per cent, fuelled by its range of aggressively priced products.

Earlier Sony Ericsson had reported that in 2007 their sales crossed 100 Million handsets.

Sony Ericsson gained market share during the quarter due to the continued success of such products as the K550 Cyber-shot™ and the W200, W300 and W580 Walkman® phones in the Americas and Europe. Although Average Selling Price (ASP) increased slightly sequentially during the quarter, as a result of the introduction of new flag-ship Walkman® and Cyber-shot™ phones such as the W910 and K850 models, the trend for falling ASPs year-on-year reflects the company’s direction to broaden its product portfolio.

During the fourth quarter Sony Ericsson announced that it had entered into a series of agreements with Motorola, Inc. whereby Motorola acquired 50% of the share capital in U.I. Holdings BV, the Dutch owner of the Swedish software company UIQ Technology AB, which was acquired by Sony Ericsson from Symbian Ltd. earlier in the year. The transaction was ratified by the appropriate competition authorities during the quarter.

Our friend Tomi has a very good analysis on this topic on his blog here.

LTE Architecture: Flat or Not so Flat?

'Migrating to Flatter, All-IP Wireless Networks' claims this article in converge digest. People have been talking about this Flat architectures for some time now and I decided that it was time i clear my understanding on this.

While searching my library of infinite resources i finally hit the jackpot. Qualcomm presentation from LTE 2007 has an answer.

The flatness of an access network can be measured by the depth of its link layer-specific network element hierarchy.

Going back to the article mentioned earlier:

Despite the growth of carrier networks and the evolution of standards, voice and data communications have not evolved in synch. Carriers have historically added data communications as an afterthought to voice network architectures originally conceived in the circuit-switched era, resulting in complex hierarchical networks that support both voice and data.

This type of architecture is expensive, leading to high operating and capital expenditures for service providers and significantly lowering margins in a highly competitive industry. In addition, networks employing cobbled-together voice and data communications systems simply do not have the capacity to provide the rich multimedia services and omnipresent Internet access that today’s wireless customers demand.

To address these limitations, service providers are moving toward emerging all-IP wireless technologies that promise to reduce complexity, simplify the wireless core, and decrease service providers’ operational and capital expenses.

Currently there are several initiatives that operators are considering for building wireless IP networks:

· WiMAX End-to-End Network Systems Architecture: Defined by the WiMAX Forum Network Working Group (NWG) and leveraging the IEEE 802.16e WiMAX interface.

· Long Term Evolution (LTE): Being defined by the Third Generation Partnership Project (3GPP) and targeted as a successor to GSM-based technologies.
· Ultra Mobile Broadband (UMB): Being defined by the Third Generation Partnership Project 2 (3GPP2) and targeted as a successor to CDMA-based technologies.

All three of these architectures are similar in that they leverage a flat, user-plane, all-IP network architecture with fewer nodes that enables mobile operators to integrate the core with the access network, providing real-time multimedia and broadband IP services from the core to the mobile station. This flatter architecture results in reduced latencies and thus optimizes performance for real-time services such as voice and video.

So going back to the Qualcomm presentation and checking if the LTE part is as flat as claimed.

We can see that the LTE Network Architecture is Lumpy rather than flat. Even though it is an improvement from the Release 99 (or rather Release 6) its not as flat as claimed.
Sure this would be something to consider in case of 4G (IMT-Advanced).

Tuesday, 22 January 2008

2008: Which Technologies?

There are many tecchnologies being standardised or about to be rolled out this year. A summary of these along with the predictions are available on RCR News website:
WiMAX Certifications: As promised, the WiMAX Forum opened up its labs in December for mobile WiMAX certification testing. The expectation is that 802.16e-based products will start earning the forum’s stamp of approval by mid-year with, “hundreds of devices to go through certification,” over the course of the year. The larger hope is that certification will drive operators currently testing the technology (300-plus by some estimates) into broader, commercial deployments.

In reality, certification may be little more than a formality. Why?

1) Operators wouldn’t be deploying kit that they didn’t think complied (at least somewhat) to the 802.16e standard and WiMAX Forum profiles.

2) The forum’s plugfests have helped to get vendors roughly in-line in terms standards implementations.

3) Even if you don’t agree with points one and two, consider the following: Some so-called mobile WiMAX products may never get certified. Implementations in potentially popular bands such as 5 GHz or 1.5 GHz may always remain relative niches, never garnering enough interest from vendors to warrant the creation of WiMAX profiles in these bands. And, while Taiwan is oft painted as a hotbed of WiMAX device development, vendors there are reportedly bristling at the prospect of costly certification testing, setting the stage for certification delays or two tiers of products (fully certified and compliant, but not certified). Now, that would be buzz-worthy.

Long Term Evolution: If only in trade shows, WiMAX and LTE seem inextricably linked — usually separated by the word, “versus.” And, as with WiMAX, the LTE market will steadily develop throughout 2008. Tradeshow booths will be filled with prototype LTE systems. The 3GPP should make progress on getting the standard completed. Trials (based on finalized radio interface specs) will take place. Additional operators will declare their allegiance to 4G. Yep, it will be a good year for LTE … and we’ll hear about it often. None of these trials or demos, however, will change the facts. Commercial kit won’t be available until sometime next year; until then, we won’t know how any vendor’s solutions actually perform. Broader field trials will likely be a 2010 phenomenon given the habit of device availability trailing networks by a good margin. This means that commercial service deployments might be expected in the 2011 timeframe. Considering the fact that operators will continue to ramp up 3G usage, that might even be optimistic. So, at the end of the day, enjoy the show floor demos, pick up a few flashing toys for your kids at home and know this: The LTE landscape at the end of 2008 should look a lot like it did at the end of 2007. LTE will be broadly accepted as a dominant 4G standard, but services will be years away and real-world performance will be a question mark.

Note: On referring LTE to 4G please read this and this.

700 MHz spectrum: With all due respect to my troglodytic (cave-dwelling for those didn’t get My Word Coach for your Nintendo this Christmas) friends, the FCC’s 700 MHz auction begins later this month. In the run-up, intrigue around who will bid, how much they will bid, and the bidding/service rules they will face has provided fodder for nearly daily news. Once the bidding begins, we’ll get day by day (and hour by hour) updates on the auction process and what it means for the U.S. telecom landscape. I personally plan to avoid the fray by being in the middle of the ocean when the auctions start and then retreat to the West Virginia border (true story, I go from vacation to our annual corporate retreat at The Homestead). The buzz, however, is somewhat understandable, given the propagation characteristics at 700 MHz (great for broad coverage and in-building penetration), and the fact that it could support new market entrants … not to mention the potential for 700 MHz developing into globally standardized wireless spectrum.

But, what’s going to happen this year? Auctions will take place and spectrum will be allocated. Networks, however, won’t get deployed; I’m planning to continue watching analog TV with my old rabbit ears right up until I can’t anymore. More importantly, operators looking to 700 MHz as a platform for their 4G networks may need to wait for several years before they can even get access to the equipment they need (see rant above).

And, the value of propagation and global standardization? The prospect of 700 MHz spectrum being available around the globe will doubtless bid up the value of the band here in the U.S. and price me out of winning the B-Block CMA covering my house.

Yet, it will take years, if ever, before other countries can move on the band. By that point the costs of developing multi-band devices should be lower, perhaps thanks to innovations like software-defined transceivers. At that point, any operator building out a 4G network should be building for capacity and FMC solutions leveraging Wi-Fi or femtocells (or something we haven’t yet heard of) should solve most of our in-building coverage problems (cue the anti-femtocell rhetoric).

Femtocells: Speaking of femtocells … I like femtocells. I even have one on my desk; it’s a mostly-empty mock up handed to me last year as one of the industry’s pioneers was looking to empty out their booth at the end of 3GSM (back when it was still called 3GSM). 2008 should be a banner year for the little boxes since trials will start taking place beyond the more limited moves in 2007. These trials, in turn, will be critical for providing insights into how (or if) the products work. Just as we’ve heard about most major femtocell RFPs and commitments, we’ll hear about these trials as they move forward (often thanks to tradeshow sessions). What these trials won’t do is provide any immediate answers.

Yep, 2008 will be a year of figuring out what makes sense: the best business models, the best mobile core integration options, the best device management strategies, the best interference mitigation strategies, the reality of zero-touch installations, the tradeoff between products filled with bells and/or whistles and a bill-of-materials that can support profitable services. To this end, Sprint Nextel deserves credit for getting the ball rolling early, deploying commercial services instead of waiting for others to figure it all out. Yet, for all of the femtocell buzz (some of it warranted), 2008 will be largely about setting the stage for 2009.
More of it here.

3GPP LTE Progressing Nice and Steady

Press release from 3G Americas yesterday, applauds the progress of LTE standardisation over the recent months.

“3G Americas has consistently supported the 3GPP standards for the GSM family of technologies and we applaud the progress of the LTE standard. LTE is a dynamic OFDM-based next generation technology with a great future. 3G Americas is committed to not only today’s global leading standards of GSM, EDGE and UMTS/HSPA but also to the innovation that LTE provides our industry,” stated Chris Pearson, President of 3G Americas.

Pearson added, “With this recent approval, the LTE Terrestrial Radio Access Network technology specifications will be under change control, and included in the forthcoming 3GPP Release 8.”

Meanwhile Martin suggests in his blog that voice will be the real test for LTE networks.

While on mobile Internet devices it might be acceptable to have Skype or some other VoIP technology, smaller handsets who's main purpose is voice service might have a much more difficult time with VoIP. Such devices must also be capable of roaming to 2G and 3G networks when running out of LTE coverage. Especially in 2G networks, VoIP is a no go so voice service must fall back to good old circuit switched telephony. So apart from the question of how VoIP will be done on LTE handsets the much bigger question is how to make the experience seamless over 2G, 3G and LTE.

More about that here.

Tuesday, 15 January 2008

80 million mobile WiMAX subcribers (by 2013)

A new report by Juniper Research calculates the number of mobile WIMAX subscribers will exceed 80 million by 2013. The biggest surge in growth, says the research firm, will happen after 2010. Juniper’s projections assume a wide range of attractive devices will be available on the market within three years (at competitive prices), and mobile WiMAX operators will achieve service differentiation from mobile operators.

According to figures from the WiMAX Forum, WiMAX technology had the potential to reach 2.7 billion people before the ITU announcement. That number now rises to over 4 billion.

There are already some interesting announccements regarding WiMAX in the last one month:

Milton Keynes Council has launched what is thought to be the UK's first commercial wireless broadband service using WiMAX technology. ConnectMK, a private company set up by the council to address the issue of poor broadband connectivity across Milton Keynes, has joined forces with Freedom4 to provide residents and businesses in the area with access to WiMAX services.

For those who dont know, Milton Keynes is a relatively newly developed town in Greater London area. When it was being expanded in 1980's, the engineers decided they can save lots of money by having copper plated aluminium cables rather than copper cables for telephone, etc. Their experiment was successful and received lots of applause untill the arrival of ADSL when people realised that these cables cant be used for carrying broadband ;)

Sprint said that the Xohm service will be commercially available in select cities around the United States in the second quarter of 2008.

Sprint was one of a number of vendors at the Consumer Electronics Show with big WiMAX plans. For its part, Sprint said that the Xohm service will be commercially available in select cities around the United States in the second quarter of 2008. The company hopes for a large-scale rollout of its Xohm WiMAX service by the end of the year.

Other CES vendors with WiMAX-related announcements: San Francisco's OQO (an ultra-mobile PC planned for 2008); AsusTek of Taiwan (a variety of WiMAX-embedded devices); and Zyxel (collaborating with Sequans on WiMAX access devices for Xohm's commercial launch).

One of the promises of WiMax, a service Sprint will be providing under the Xohm brand, is that receivers for it can be built into a variety of devices like cameras and Web tablets that usually don't have a built-in Internet connection or rely on Wi-Fi, a short-range technology.

By the way, it seems Sprint is already going through a rough patch:

Now, another round of subscriber losses is expected for the fourth quarter ended Dec. 31 as the firm tightened credit standards for would-be subscribers and failed to price its handsets competitively, said Philip Cusick, an equity research analyst with Bear Stearns in New York.

Shares of Sprint, which have lost more than 25 percent in the past year, closed up 11 cents yesterday at $12.36.

Citing anonymous sources, the Wall Street Journal yesterday reported that Mr. Hesse plans to fire several thousand workers. The company last year fired 5,000. The Journal also disclosed that Mr. Hesse may relocate the company's headquarters to Overland Park, Kan., where 13,000 of the company's 60,000 employees are located.

And finally the rest:

Airspan Announces 5 GHz WiMAX Deployments in Ukraine

Alcatel-Lucent opens WiMAX interoperability testing center in Taiwan

Kirkland, Wash.-based Clearwire, founded in 2003, is building a nationwide high-speed wireless network based on WiMax technology. Under the agreement, Clearwire would offer Google's email and calendar applications to its customers. In the future, Clearwire also plans to offer Google's search tools.

Chrysler to put WiMax into its cars

Axtel (Mexico) selects Motorola for WiMAX deployment

Alvarion and Bridgewater Team for Joint WiMAX Deployment in Angola

Friday, 11 January 2008

I-HSPA: HSPA+ by another name

Nokia refers to HSPA+ by the name I-HSPA or Internet-HSPA.

According to their whitepaper:
  • 3G operators who have deployed I-HSPA have flat network architecture similar to LTE/SAE in place, and can thus cost-efficiently introduce LTE/SAE.
  • 3G operators with a deployed WCDMA/HSPA network can migrate
    directly to LTE/SAE. Migrating to the flat network architecture of
    Internet High Speed Packet Access (I-HSPA) may also be beneficial
    because it accommodates LTE/SAE’s flat IP-based network architecture while supporting legacy WCDMA/HSPA handsets. The operator can thus enjoy the transport and network scaling benefits immediately and easily upgrade the network to LTE/SAE later.
  • Greenfield and CDMA operators can introduce LTE/SAE networks
    directly or follow one of the above paths. GSM/EDGE may be a good choice for strategies more immediately focused on voice centric business. Operators opting to take the I-HSPA path can capitalize on the ecosystem of HSPA terminals, benefit from the flat architecture today, and quickly optimize mobile broadband performance.

Heard of MobileFi?

A new book title came to my attention, "Wimax/Mobilefi". I hadnt heard of MobileFi before so i did a bit of digging up and here are the details.

MobileFi is also known as Mobile-Fi or Mobile Broadband Wireless Access (MBWA) and is better known as IEEE 802.20

IEEE 802.20, also referred to as Mobile-Fi, is optimized for IP and roaming in high-speed mobile environments. This standard is poised to fully mobilize IP, opening up major new data markets beyond the more circuit-centric 2.5G and 3G cellular standards. The Mobile Broadband Wireless Access (MBWA) Working Group was established as IEEE 802.20 in December 2002. Its main mission is to develop the specification for an efficient packet-based air interface optimized for the transport of IP-based services. The goal is to enable global deployment of low-cost, ubiquitous, interoperable, and always-on multivendor mobile broadband wireless access networks. IEEE 802.20 has designed a new physical layer (Layer 1 protocol) and MAC/link layer (Layer 2 protocol) around IP packet Layer 3. It can operate in licensed bands below 3.5GHz, with cell ranges of 9 miles (15 km) or more. IEEE 802.20 can operate at speeds of up to 155 mph (250 kph).

Unlike WiMAX, which was incubated inside IEEE 802.16 family and evolved from earlier 802.16 technologies, 802.20 [5] or Mobile-Fi was designed from ground up as a technology to support high-mobility services. It aims to support mobility as high as 250 km/h and a peak rate of up to 260 Mbps in the licensed spectrum below 3.5 GHz. Th e enabling technologies are also OFDM, MIMO, and beam-forming. The draft standard is still under the IEEE standardization process.

Both WiMAX and 802.20 use OFDM-MIMO, which is emerging as the main technology for future cellular packet data networks, including 3GPP long-term evolution and 3GPP2 air interface evolution as well.
While the data rate and range of Mobile-Fi are only half those of Mobile WiMax, Mobile-Fi is inherently more mobile. It has an astonishing latency of just 10 milliseconds (500 milliseconds is standard for 3G communications) and can maintain integrity at speeds as high as 155 mph (250 kph), compared to just 60 mph (100 kph) for WiMax. Because it uses more common spectrumlicensed bands up to 3.5GHzit also offers global mobility, handoff, and roaming support. Whereas Mobile WiMax is looking at the mobile user walking around with a PDA or laptop, Mobile-Fi addresses high-speed mobility issues. One key difference is the manner in which the two standards are deployed. One assumption is that the carriers are going to deploy Mobile WiMax in their existing (802.16a) footprint as opposed to deploying a more widespread footprint, like a cellular network. Because Mobile-Fi is aimed at more ubiquitous coverage, a larger footprint will be required.

Countries and companies often seek to control the market by developing standards they hope will dominate the global scene. The United States has led the way with IEEE standards, and the European Union's ETSI standards are their counterparts. The work of standards consensus is ongoing, uncertain, and difficult to predict. Mobile operators, who are generally friendly to Mobile WiMax, see Mobile-Fi as a competing standard that could make their 3G licenses worth rather less than they paid for them. The fact that Intel is behind WiMax is a strong force and will undoubtedly push the WiMax standards forward.

Mobile-Fi will have to overcome several hurdles. First among them is the fact that it can be used only in licensed bands below 3.5GHz. Another is that Mobile-Fi trails the Mobile WiMax standards process by a couple years. Another hurdle is whether there is indeed a large requirement for 155 mph (250 kph) handoff. In addition, we do not know what effect Mobile WiMax being nationalized in Korea will have. And, very importantly, cellular companies may not be willing to undercut their 3G service. Certainly, we can assume that the US$100 billion investment in 3G spectrum by the European mobile carriers alone might be weighed against a workable Mobile-Fi standard. With the possibility of proprietary systems (e.g., WiBro, Flash-OFDM) being in place a number of years before Mobile-Fi is standardized, the likelihood is that by then, Mobile WiMax will be backward compatible with WiMax fixed services. Licensed or unlicensed, Mobile-Fi will not be ubiquitous, and WiMax probably will.

Further Reading:
* Book: Wimax/Mobilefi
* Official IEEE 802.20 website

Friday, 4 January 2008

HSPA Data Rates Calculation

People often get lost while calculating the data rates for HSDPA, HSUPA or HSPA+

Note: HSPA+ is better known as eHSPA or HSPAe where e stands for evolution or evolved

Most people are aware that the theoretical maximum for HSDPA is 14.4Mbps, so lets see how we reach 14.4Mbps:

In UMTS, in each slot the maximum number of bits transmitted is 2560. The correct term to use is chips rather than bits. If you want to know where this 2560 comes from or why chips then please refer 3GPP TS 25.211

There are 15 slots per 10ms frame so since the TTI for HSDPA is 2ms, there will be 3 slots. So there will be a total of 7680 chips.

QPSK has 2 bits per symbol = 7680 * 2 chips for 2ms = 15360 chips/2ms = 15360 * 1000 /2 chips per second

Now the SF is fixed at 16
= (15360 * 1000) / (2 * 16)
= 480 Kbps

Terminal that uses 15 QPSK codes will get 480 * 15 = 7.2Mbps

On other hand 16 QAM will have 4 bits per symbol so the rate would be 7.2 * 2 = 14.4Mbps

In HSPA+ we will also have 64QAM which has 6 bits per symbol (2^6 = 64) so the max rate would be 7.2 * 3 = 21.6Mbps.
The figure above is self explanatory and shows the data rate in case of eHSPA.

Thursday, 3 January 2008

Ericsson Networks to launch MBMS in March 2008

Ericsson Networks last month successfully completed IOT (Inter-operability Testing) with EMP (Ericsson Mobile Platforms ... Handset people) and LG.

The plan is to have early launch systems in March 2008 followed by full deployment in September 2008. The only limiting factor might be the handsets which are not ready yet and have many problems that need to be sorted out.

According to CIO, Australians might be the first to get their hands on MBMS networks.

From what I have seen, Nokia is much ahead of its rivals in the MBMS game but they are keeping mum on this topic ... atleast for the moment. Sure it will be an interesting 2008.

Happy New Year

Its still not too late ... HAPPY New Year to everyone.