7 October 2011

NetEvents EMEA Summit Frascati - Press and Vendor meetings summary.

Vendor meetings summary:

A key element of the NetEvents press summit formula is the meetings between press and vendors, so here we present a summary of some of the vendor meetings.

ECI Telecom - Matthias Nass, VP Field Marketing, EMEA

About the company: we have 2,600 employees, we do internet transport and wireless infrastructure, BT is a key customer and uses our equipment for next generation access. We are strong in utilities - especially those moving towards becoming a u-telco. We are strong in India, China and Latin America. We see telcos moving to the cloud - such as Deutsche Telekom's T-Cloud service.

Vision: we need more fibre to the kerb - or at least closer to the end user - if cloud is to succeed. What do telcos need to become successful? End user acceptance. Today, SMBs are quickest to the cloud. Telcos need to move away from the network and add more value.

Yet telcos are focusing on big enterprises not SMBs. They are slower to the cloud because they have problems of regulation and net neutrality that SPs do not. Russia and Asia have fewer regulatory brakes so end users are getting fibre where Europeans are not - but problem there is multiple providers aiming for the richest / low-hanging fruit.

VMware - Joe Baguley, Chief Cloud Technologist, EMEA

VMware recently made a vCloud global connect announcement where three service providers announced a global cloud: BlueLock, SingTel, and COLT. The aim is to allow you to move VMs easily to an service provider in the cloud. Some telcos provide secure/ guaranteed bandwidth which they can then put a service wrapper around as Orange Business Services does - it sells cloud as a service: you don't rent servers but a service.

People don't want to manage operating systems so telcos and service providers will move away from infrastructure as a service (IaaS). Rather, people want managed services, for example CloudFoundry, which is an orchestration a layer above a bunch of virtual machines (VMs) and allows the hardware platform to provide the resilience etc.

We need to learn how to orchestrate VMs - you can then sell it to enterprises and service providers. Everyone is moving up the stack from IaaS upwards. CloudFoundry is doing this - is the biggest growing part of cloud and telcos will move there.

The future will see MVP - mobile virtual iPhone / Android - which allows you to deploy a work phone onto your Android, which is good for BYOD (bring your own device) environments. Mobile providers and telcos need to understand how this works as there's only one SIM. We are talking to providers about that - they are arguing and bidding as to who can be first to do it. Samsung and LG are first hardware vendors involved.

Other news

Check out VMware AppBlast - delivery of apps to any html5 device.

We bought SlideRocket, Zimbra - have Project Octopus (from EMC Mozy) - is a corporate Dropbox that does collaborative documents, chat etc.

VXLan - also launched at the recent VMworld - can move VMs around without breaking VLANs and IP addresses.

Tata Communications - James Walker, Head of Managed Network Services

We are working with global companies just below the top tier. We aim to be more flexible, more aggressive, eg selling network services on a per-seat basis. Companies that do a lot of M&A are interested. It means the customer no longer has to dictate the network topology - we do all that.

We launched a virtual Ethernet platform last year that allows multiple datacentres to be treated as one virtual datacentre, good for companies spread across multiple facilities. It allows us to do deterministic routes.

This is all on our own network - we own about 25% of the world's capacity - large proportion is used by our own IP business, about 2.5Pbits/month, plus clear channel services which adds up to about 80% of the total bandwidth, while the rest is consumed by MPLS and voice. We will shortly own a ring of fibre all the way around the planet.

Datacentres - we bought TeleGlobe and Tyco Networks so we inherited those, and built 15,000 sq.m. datacentre facility in Singapore, and others in India and S Africa, all on our network.

This supports our new InstaCompute scalable cloud offering in Singapore, and covering neighbouring countries: Malaysia, Hong Kong, Thailand, Indonesia, Vietnam and the Philippines. The IaaS service enables Asia Pacific businesses to use a global infrastructure via the Internet without the cost and complexities associated with IT provisioning.

Tata aims to be in top five of all the businesses we are in. Competitors operate in Europe and are unsuccessful but we have a $100m business here - we sell well in France.

Mobile is not strategic for us - we are focused on fixed line.

There's a gap in the market for mobile experience because in developed nations it's very computer-centric not mobile-centric.

In Africa, mobile backhaul is going mostly satellite based using Google's upcoming EchoEye satellite.

Verne Global - Jeff Monroe, CEO; Tate Cantrell, CTO; Lisa Rhodes, VP Sales/Marketing

JM: We are bringing dual power 500 sq.m. datacentre to Keflavik, Iceland - it arrives in October. We covered negotiations on tax policy, power utilities, running cables to Iceland. First customer is DataPipe - services US and UK with management and cloud services. The cable was laid in 2009. This is the world's first mission critical data centre with zero-emissions to be manufactured and constructed in less than four months.

We got Icelandic tax law changed so that VAT could be refunded if equipment was deployed - ie this is in line with the rest of Europe. We had to install power and telecoms lines. The site is former NATO air base. Dual power source is geothermal and hydro - but we have diesel generators as backup - customers demand it just in case the Earth stops generating heat! Iceland has a very robust electricity supply because there is constant demand: there are a lot of aluminium smelters using big power - they use 85% of the total constantly so the whole varies little, unlike eg the UK.

DataPipe offers green services. We have multi-terabit cables that deliver point-to-point connectivity.

Customers save money because we don't add a green premium. We are a wholesale datacentre provider. We use COLT modular datacentres as the basis. We looked at lots of other modular datacentres and the plus point for COLT was our ability to collaborate.

There is not much waste heat: power cost and environmental impact are zero. So is PUE important? We quote the PUE to customers - have not yet announced it. we can do better than the COLT PUE: 1.21. The main energy cost is humidification, and inside the box temperature will be maintained to ASHRAE standards.

Do people care about green in a recession? There has to be a cash advantage, so the renewable part of this is a key advantage for us.

This is a 40-50 year investment so we think about things in long timescales. We know what the cost of electricity will be in 20 years time - no-one else can do that. The price we pay for electricity is under 50% - all other things being equal - compared to UK power cost.

Wellcome Trust is the majority investor.

About COLT: we see a strong demand in the colocation market and we required a partner who could provide highly resilient, flexible data centre space, configured to our specific technical requirements.

Infoblox - Steve Garrison, VP Marketing; Francisca Calkoen, EMEA Marketing

Preamble: the need for IP address management is growing as rates of change and the numbers of IP addresses increase. It means there's a need for automation, driven by trust of the system. (we do DDI=DNS, DHCP, IPAM) - it's a network provisioning control plane tool.

Our systems run as a grid so that they understand what's happening using a distributed database. It's a resilient network fabric. It means we can segment network services and the grid can get as big as you like.

New: the new product is Multi-Grid Master (MGM). You can install 50 grids, each of which has 250 devices = 12,500 devices. It's a virtual mainframe... Price: $25,000.

Brocade - John McHugh, Chief Marketing Officer

Customers are worrying about economic issues and technology ageing. Looking at Cisco's VCE - it's a Ford Model T. It achieved success through being limited but customisation came along. VCE is fine for a first step but it needs to be customised.

Brocade is about CloudPlex - the best systems (storage, computing, networking) need to get together. We worked with Avnet to build an integrated stack with twice the performance of VCE at the same price, with flexibility and an open architecture.

Cloud moves IT to opex not capex - IT used to be like building your own power plant but now you don't need to do that, you can acquire applications etc on a per-use model. This is transformational. You can even buy your networking on a subscription model with Brocade Network Subscription, announced at VMworld four weeks ago. This is appealing to those in the managed services business as they won't get cashflow positive until the second or third year. This is where resellers are going as it means they don't have capital charges. the proof is that MSPs are pressing other vendors to do likewise.

We designed Ethernet fabric chips after the acquisition of Foundry Networks, so our network fabric becomes a flexible hardware-based, lossless distributed switch chassis. It knows about the port configurations and topology. It means the system is multiply redundant. This is good for internal network traffic inside the datacentre - which is not what Cisco's Nexus architecture is designed for - it's designed for north/south traffic not east/west traffic.

Brocade Ethernet Fabric called VDX - we have 200 customers (60 in EMEA) for this - not big numbers but this is datacentre stuff - equivalent to Juniper QFabric. VDX is doing well: Q211 revenue for this up 78% Q311 up 119%. The system has Fibre Channel ports so it can link into existing storage networks.

Growth of cloud will be service providers and SIs building new datacentres not enterprises.

MEF – Kevin Vachon, COO and Phil Tilley, Global Marketing Co-chair

Kevin presented an impressive set of market data and predictions for further growth in Carrier Ethernet (CE) equipment and services. Vertical Systems Group’s recent report confirmed that the worldwide market for Business Ethernet services is on target to reach $44.3B by 2015 with double-digit global growth despite recent economic uncertainty.

Having just celebrated its 10th anniversary, the MEF continues to gain momentum through its increasing membership and other initiatives, including an expansion of its successful certification program to include MEF – CECP’s – Carrier Ethernet Certified Professionals important new standards wok in the areas of access services and 4G mobile backhaul. MEF’s strategy for fully enabling Carrier Ethernet to support Private Cloud Computing also forms a major component of MEF’s current work plans.

The MEF is on track with important technical and educational work that will enable mobile operators to more easily implement MEF’s widely adopted Class of Service models in wireless backhaul environments, to drive down mobile backhaul costs. Also, with work nearing completion on the next generation of MEF’s Mobile Backhaul Implementation Agreement, mobile operators will be have the necessary technical guidance to now support both 3G and 4G.

Building on the Global Interconnect Program of 2010/11 significantly, MEF will be introducing a new access service specification in early 2012 that will standardize wholesale access service implementations and enable shortened cycles for the sale and provisioning of such services. This new area of work is seen as one of the final major puzzle pieces in support of Carrier Ethernet globalization.

No comments: