21-05-15

IP Expo Manchester – Day 2 Review

Today was the second and final day of the inaugural IP Expo event at G-Mex Manchester Central. I’d spent a lot of time wandering around the solution hall yesterday, so today I wanted to spend more time in the sponsor sessions. The problem is that often I would find myself bumping into a familiar face and at the end of a catch up, 30 minutes has gone!

One thing I’d like to express before I go through the session reports is that I wish vendors would top this “avoid lock in” bullshit. Every vendor does lock in of one form or another. After all, how can you expect a good revenue stream for a product if it’s too easy for a customer to take their stuff elsewhere? I’m not saying vendor lock in is a bad thing, as long as there is a degree of interoperability which means you do have a path out if you need it. But please, don’t sell me something proprietary and then tell me it avoids vendor lock in, it makes my blood boil.

Anyway, onto the show…

Microsoft Mobile First, Cloud First Transformation – James Akrigg

Much to my disappointment, I missed all of this session bar the last 10 minutes. I’ll not go into the details as to why, but some aspects of the event organisation could be filed under “could do better”. Not all bad though. Anyway, I caught some integration with Cortana and Business Intelligence. Impressive stuff and makes you realise how much consumer tech is now appearing in “business” products as users start to expect the same experience on both sides.

One thing I will say about Microsoft is that I’ve been pretty impressed with how they’ve reinvented themselves. It’s not easy to change the path of the company that is so big and has massive cash cows like Windows and Office. In many ways, it makes me wonder how much further along they’d be today if they’d ditched Steve Ballmer a lot sooner.

The full house signs also demonstrated that Microsoft is still more than relevant in today’s IT landscape. Reports of their demise have been premature, and hats off to them for reinventing themselves as a cloud/mobile company. Yes, there are a lot of updates about Windows 10 and Office 2016, but they seem to me to be on a par with Azure and other platform announcements these days.

The true foundation for the Software Defined Enterprise – John J. Ryan – VMware

jr

John Ryan from VMware – not walking like an Egyptian..

Next up was John Ryan from VMware to take us through a “how we got here” session and also how VMware were driving the SDDC market forward.

Key points included:-

  • We’re now in the mobile/cloud era after mainframe and client server eras
  • Cloud management platform is automation, operations,  business intelligence wrapping over “traditional” virtual infrastructure of compute, storage and  networking
  • Control of data centre automated by software (management and security)
  • Foundation of Software Defined Enterprise means handling complex tasks in a simple way
  • vSphere provides capability to virtualise applications, desktops, servers, databases etc, best general purpose hypervisor around
  • Hadoop Big Data extensions, certified support for SAP HANA amongst newer features
  • Container support in vSphere 6, integrated OpenStack
  • Instant clone for desktop workloads, radically improve your VDI provisioning
  • 4x scalability in vSphere 6 (vCPU, vRAM, etc.)
  • Photon special VMware edition of Linux to run containers, open sourced with Lightwave (identity)
  • Long distance vMotion up to 150 milliseconds latency
  • Enabled by cross vCenter vMotion, which is also new in vSphere 6
  • Use cases for long distance vMotion include follow the sun, disaster avoidance
  • Fault tolerance now up to 4 vCPU s
  • Needs 10 gig infrastructure however
  • Content library can store templates, ISOs,  OVAs.  Subscribe and replicate content (bit like System Center)
  • Recent tasks and right click improved in Web client, more intuitive workflows (i.e. a bit more like the “fat” client!)
  • Virtual SAN extended to use hybrid models, including storage arrays
  • Virtual volumes (VVols) changes the storage paradigm
  • Automated policy management
  • All industry partners will support virtual volumes, some natively in the array and some via virtual appliances with VASA

One cloud, One security – VMware / Trend Micro – Peter Bury and Stephen Porter

peterbury

Peter Bury – VMware

From the previous session I hot footed it across the hall to the next session which mainly dealt with what NSX is at a high level and how that fits into the SDDC message. This was a decent presentation that illustrated the issue of applying security out on physical network devices when you might want to be able to segment VMs away from each other in the same cluster and the “old” networking model created in vSphere with standard vSwitches just wasn’t very flexible.

Key points from this session:-

  • Old design methodology meant sending VM traffic out of cluster and let edge firewall deal with it (tromboning)
  • Firewall rules get massive over time as nobody removes them when a service is unprovisioned because of the fear of breaking something. The secondary impact is slowing down the firewall as it has to churn through 100s of rules
  • Not a fluid design for agile changes
  • This design is virtual but it’s not cloud
  • Enterprises want rapid elasticity, roll out services as the organisation demands
  • If IT is too slow, company will go out to the public cloud
  • Intelligence is built into software for security, firewall etc. Physical networks become “dumb carriers”
  • Load balancing, routing, switching, firewalls, access control lists in software, as part of the hypervisor stack.
  • Context in workloads is achieved by baking in features into the hypervisor
  • New model enables wrapper around workloads to provide security
  • API allows trusted partners to provide their expertise in the micro segmentation of virtual machines
  • Anti malware,  anti virus, intrusion protection etc
  • Agent less design from Trend
  • All policy driven so policy follows VM wherever it goes
  • Moved from scheduled scan to real time security
  • Dashboard available for vCenter/vRealise Operations Manager
  • Security rules can be applied before patches are ready. Heart bleed rule available in a couple of hours, patches for the same vulnerability take days or weeks

It’s time to upgrade from backup to business continuity – Fifosys

fifosys

The next session covered the topic of business continuity. One interesting point on this was that BC is not always a “site down” issue. Sometimes it can be a key LOB application that has gone for a lie down and you need to have a strategy for bringing that back so the business can keep functioning properly.

This was also a leader into introducing us to the Datto appliance, which I must admit was new to me. There seems to be a burgeoning market now in hardware appliances that keep some data local and then move “cold” blocks out to the cloud where it’s cheaper to store. This product works along similar lines, but with backup images.

Key takeaways:-

  • Disaster recovery hampered by slow,  manual processes including getting off site backups,  restore from tape etc
  • 13 % of Fifosys survey responders don’t take tapes off site
  • 61% still using tape
  • 52% of SMB s do not have a BC plan as they don’t view IT as critical to the business
  • 45% of downtime caused by human error according to Oracle User Group
  • BC is not just failure of a site, but key business systems
  • Assign a financial cost to an outage to justify a BC plan, there are plenty of simple equations out there you can use
  • Impact time has a direct effect on the costs of an outage
  • BC should be simple and automated
  • Sub 1 hour recovery is a must
  • Reduce reliance on staff for BC
  • BC reduces operating risk, don’t see it as a cash waste
  • DR tests should not be time consuming or impactful
  • Traditional designs use active/passive or active/active data centres, which can be prohibitively expensive
  • Datto appliance performs backups and replicates to the cloud
  • Image based backup every 5 minutes
  • Can restore to the Datto appliance as it has a baked in hypervisor
  • Agent based backup
  • Screen shot verification is an automated daily DR test
  • Uses inverse chain technology
  • Datto protects 100 PB globally

Cisco’s Intercloud Strategy – Bruno Oliveira

cisco

Cisco have a “cloud of cloud” strategy called Intercloud, which is an interesting concept. By the presenter’s own admission, there are still odd pieces of it not quite ready yet (mainly the Cloud Market Place option) but again in a similar way to vCAC/vRA does it’s best to be a technology agnostic solution (with Cisco’s wrapper around it, naturally!).

Again the key takeaway is freedom of choice and the flexibility to move workloads around to internal or external clouds as economics and performance requirements dictate.

Key points:-

  • Uber now biggest taxi company in the world but don’t own taxis
  • AirBnB don’t own hotels
  • Digital disruption caused by these types of companies
  • 50 billion objects connected to the Internet by 2020
  • Unified workload management. Any VM any cloud is the essence of Intercloud
  • Keep data in country as opposed to AWS etc where you may not know where it is
  • 55% companies turning to the cloud to lower costs
  • Global cloud of clouds using VMware,  OpenStack, etc. Cisco validated architecture
  • Cisco want to wrap around all these disparate services so customer sees it as their infrastructure
  • 160 inter cloud partners
  • 60 providers and resellers
  • InterCloud fabric is the software wrapper around this environment
  • InterCloud market place coming,  both for internal and external clouds
  • Fabric provides end user and administration portals
  • Can move workloads from one provider to another (Azure to vCloud Air, vCA to AWS, etc.)
  • Cloud usage collector can be attached to physical network kit to accurately measure cloud service consumption so CIOs can “really” see what external services are being consumed

A new approach to optimising the WAN with Citrix CloudBridge – Al Taylor, CloudDNA

My final session of the conference was around the Citrix CloudBridge 11 appliance. Folks who know me know I’m not so much a Citrix guy, but I try to be as agnostic as possible and try to avoid “drinking the Kool-Aid”. At the end of the day, I don’t believe it pays to close yourself off from any vendor as you never know when they’ll have a niche product or solution that will come to your rescue.

I actually enjoyed this session the most of all the ones I saw over the two days. There was something vaguely punky about the presenter and his enthusiasm for the CloudBridge device really shone, whereas some of the other presenters went through the motions a little bit. I’m not denigrating them, but perhaps that’s the difference between a true tecchie and non-tecchie speaker.

Anyway, CloudBridge is a Citrix appliance that is intended for use over constrained bandwidth to improve XenDesktop / XenApp user experience (amongst other use cases such as video and Lync).

Key session points:-

  • CloudDNA are the only dedicated Citrix cloud networking practice in the UK
  • NetScalertaylor.com for NetScaler blog
  • ILoveNetScaler.com news aggregate and weekly newsletter
  • CloudBridge is like WAN repeater
  • Acceleration and compression, amongst other things
  • Video optimisation for Lync etc
  • HDX analysis for CloudBridge to get full visibility of all bottlenecks
  • Feeds back into Desktop Director for quick and simple performance analysis
  • 64 channels in ICA traffic
  • Prioritise channel traffic to ensure performance for the user
  • Tolly Report on CloudBridge gives WAN optimisation report of optimised vs non-optimised
  • Virtual appliance or piece of hardware
  • Branch office in a box, can run ThinPrint on the hardware if need be
  • Virtual WAN binds multiple links together and uses policies to decide which traffic goes down which link. Not bonded
  • Encrypt the paths between the two end points
  • Send packets based on application needs
  • Active bi directional probing
  • 600 applications optimised out of the box, not just Citrix centric
  • Faster time to deploy branch offices

In all, I enjoyed the event and it was nice to see this type of event in my neck of the woods as these things tend to be London only. Hopefully there was enough interest to make the show a bit bigger next year (they’ve already published similar dates for 2016) and get some “proper” representation from the heavy hitters (VMware, Microsoft) rather than being a desk on part of a partner’s stand.

 

20-05-15

IP Expo Manchester – Day 1 Review

As part of my new role, I’m hopefully going to be getting out and about a lot more and seeing what the virtualisation market has to offer that is complementary to our VMware offerings. As a result, I attended day 1 of the new IP Expo conference at Manchester Central today (that’s the G-MEX in old money!), where there were a whole slew of VMware partners showing off their wares. Not only VMware too, but other vendors and some quirky exhibitors (such as the Museum of Computing).

My first impressions of the solutions hall was that it was a bit smaller than I was expecting, taking up just half the hall. I suppose seeing as delegate tickets were free, the organisers wanted to dip of a bit of a toe in the water before committing to a much bigger event. The Expo also shared the venue with the TMRW conference, which is a “paid for” event with the likes of Brian Cox (that’s the scientist and not the mad Scottish actor) speaking to the masses.

Opening Keynote – Sir Richard Leese, Leader of Manchester City Council

leese

So first up was Sir Richard Leese to open the event. The general theme of the speech was to re-iterate the growth in the local economy and how that was being driven by the digital economy. There were so many facts and figures sprinkled throughout the speech that I wondered if I should have a bingo card on my lap to tick off all the numbers!

Anyway, some key information taken from the speech was:-

  • Greater Manchester population grew by 19% in the last 10 years
  • 28% growth employment rate between 2004 and 2014
  • Manchester is home to the two most popular universities in the UK
  • £2 billion invested in transport over last 5 years
  • 45,000 people working in digital economy
  • The Sharp Project in East Manchester is now fully let
  • The north west textiles industry growing fast again because of technology
  • Manchester hosts the only Internet exchange outside of London
  • Government support for “Devo Max” in Manchester area, first of it’s kind in England

At the end of the keynote we were promptly kicked out of the keynote hall for TMRW guests only. Bad form! Most of the room stood up and left, and to be honest it was all a bit awkward. I hope the organisers don’t make this mistake next year. That being said, as Tim Lovejoy was MC, I wasn’t sorry to leave!

Veeam – 3 2 1 Rule of Backups – Kevin Ridings

veeam

The next session I attended was by Veeam. I’m especially interested in this vendor as we already partner with them successfully and it’s a partnership we’re looking to extend as we grow our portfolio of solutions. Some pearls gleaned from this session:-

  • What is the 3210 rule? 3 copies of data, two different media types, one copy off site with zero errors
  • The 321 Rule protects you in most DR scenarios
  • Tape support in version 7 of Backup & Replication as customers asked for it
  • Tape provides best price per GB, doesn’t consume power to be stored
  • However, tape requires man hours to move around
  • Slow RTO
  • Veeam B&R Backup Copy jobs automatically copies backups to secondary repository, which uses different retention  policies
  • Built in WAN accelerator is aware of de-duplication and content
  • Cloud Gateway provides SSL connections over a single port
  • Additional role on the Veeam server
  • Cloud providers can securely share storage between tenants, optimising costs
  • CloudConnect works as a local repository
  • Possible to enable self service, PowerShell and REST
  • Requires version 8 update 2

The main takeaway from this session was that CloudConnect partners could be used as secondary repositories for Veeam backups and you can work with your existing Veeam partner to do this and ensure your data is safe, secure and with a trusted partner (and not “in the cloud” where you may not know where the hell it is!).

Cetus – From Ground To The Cloud
cetus
The next session was by Cetus Solutions, who discussed their project to deliver cloud services to NextGear Capital, which was a rare example of a startup company and a greenfield site. What was interesting about this case study was that as a brand new company, NextGear were not lashed to a legacy infrastructure and deliberately wanted to leverage cloud technologies to give them scale and performance without needing a large IT department and on premises kit.
  • New company founded by Cox Automotive Group in US
  • £40 million to 300 dealers in first 6 months
  • Enterprise grade infrastructure from day one
  • Highly scalable with lean back office infrastructure
  • Startup company wanted IT as OpEx rather than CapEx
  • Financial data needed to be secure and in compliance with industry legislation
  • Core infrastructure,  applications, disaster recovery as a service
  • FlexPod solution
  • Xenapp 7.5 with Atlantis ILIO
  • Citrix NetScaler with thin clients and mobile devices
  • External cloud services such as SalesForce
  • Replicate VMs to Cetus environment
  • NetScaler sat in front of desktops and environment, aggregated content from “internal” cloud and third parties into a single interface.

The key message from this case study was that this startup could be agile and keep costs and complexity down by building a bespoke environment from scratch using best of breed technologies that would allow them to scale up in future with predictable costs. Interesting!

 

SimpliVity – Making Sense of Hyper Converged Infrastructure – Hani El Qasem
hani
My next session was with SimpliVity, who are a relatively new company in the hyper converged space. I’d heard of them before owing to them sponsoring a vBeers event (hic!) but also because Cisco have partnered with them, plus I’m linked to a couple of their guys on LinkedIn and see what they’re up to.
Hani’s session was really interesting as in my mind it raised the question of whether the monolithic stacks we have now such as FlexPod and vBlock might already be a thing of the past? Yes they’re solid and reliable, but it’s still a multi vendor solution and there are a lot of moving parts. Not to mention separate compute, storage and networking takes a lot of additional overhead. I can see both sides of the argument on FlexPod type solutions and hyperconverged solutions such as SimpliVity and Nutanix. I guess it all comes down to customer requirements, as it always should be.
  • 3 1/2 years in development before shipping the product, rare for a startup
  • Best of VMworld 2013 and 2014, amongst others
  • Customers include T-Systems (aka T-Mobile), NHS
  • Customers want consumer economics on a cloud scale
  • Data protection, performance, data efficiency and global unified management all baked into the product
  • Web providers such as Facebook and Google use cheap, disposable hardware and build resiliency into the application
  • Enterprise do it the other way around
  • Converged infrastructure phase 1 like FlexPod/vBlock,  Phase 2 is storage and server. Phase 3 collapse all into one stack like SimpliVity
  • Omnicube (SimpliVity solution) is x86 architecture. Accelerator card performs dedupe,  compression,  accelerator, optimisation. Dedupe once and forever. Across all tiers (SSD, HDD, DRAM).
  • Can scale down. Minimum of 1 cube, other hyperconverged solutions need 3 as a minimum,  some need 4 if you need maintenance
  • Global unified management
  • Extension of vCenter,  not a proprietary interface. Single pane of glass plus no big learning curve of a new management UI
  • Integrated data protection and WAN optimisation
  • Reduced energy consumption, less physical space if renting CoLo space
  • One SKU
  • DRAM,  SSD and HDD tiers
  • Presentation layer uses NFS
  • 40 to 1 reduction in data storage required
  • Example given was 5.4 TB of raw data, reduced to 2.7 TB on competition solution,  650 GB on Omnicube
Virtual Volumes – Jonathan Disley Nimble Storage
After a wander around the hall, I dropped into my final session of the day by Nimble Storage. I’d lost track of time to be quite honest, so I was a little bit late and missed the first part. The session was a whistle stop tour of the new Virtual Volumes (VVols) functionality baked into vSphere 6.0. I have to say, Jonathan had to whip through it so quickly because of time that I didn’t really digest all of what was being said, so it’s something I need to follow up on. I did take some notes though:-
  • There are three main different storage designs on VMware – guest connected, VMDK on VMFS or RDM
  • Virtual volumes break into policy control plane and virtual data plane
  • Storage container is the logical construct, advertises services such as snapshot, encryption etc
  • Set storage policy and assign to a VM such as thin provisioned, snapshot etc
  • Uses VASA provider to be storage aware. Nimble embeds this into the controller. Some vendors use a separate virtual appliance
  • No VMFS partition on the virtual volume
  • Nimble InfoSight software has per VM performance monitoring. Cloud based offering that uses “call home” messages and statistics from installed controllers
  • Possible to see noisy neighbours causing latency on a data store

I’ll be back for day two tomorrow, so watch out for another review!