A Blog About Dedicated Hosting From Melbourne.co.uk

New product and service information, along with general ramblings about the web hosting industry from the Melbourne team. Also find us on Twitter!. See the live progress of our new datacentre build

SSL/TLS vulnerability discovered

November 12th, 2009

A serious vulnerability has been identified in the current SSL (Version 3) and TLS (Version 1) whereby an attacker is able to issue commands to the server that appear to be coming from a legitimate source, by exploiting a flaw in the SSL renegotiation method.

As an example, lets try the following HTTP request..;

GET /path/to/index.php HTTP/1.0
Dummy-Header: GET /index.php HTTP/1.0
Cookie: sessionCookie=Token

The good news is that although an attacker can execute the request and pass through arbitrary data, he will not be able to see the response. However, the originating client will see something different from what was initially requested.

We have done some testing with web firewalls such as modsecurity and have yet to find a way to block such requests. Looks like we will need to wait for upstream patches, particularly Apache, Microsoft and possibly a temporary release from modsecurity.

To track the exploit, see http://www.kb.cert.org/vuls/id/120541

Rob Greenwood, Technical Lead


Sales-lead solutions

November 12th, 2009

Leading on from Daniels post regarding honesty. As technical architect for Melbourne I’m a massive fan of the datacentre tours we offer both our existing and prospective clients. I think it’s a great opportunity to create close-knit relationships allowing us to:

  • Get a detailed understanding of how your business/application/website depends on its I.T.
  • Create bespoke hosting solutions designed around your requirements.
  • Give demonstrations of how a particular piece of technology will benefit you.
  • Show exactly what we’re about, and what makes us different.

A lot of the time customers come to us with quotes from competing companies. Solutions that are either not achievable, not meeting requirements or are completely overkill. It’s becoming clearer that these solutions have never seen a technical person and are being glued together by sales staff.

These companies lack the ethics to have technical people design technical solutions. It’s becoming more about the glossy PDF’s, high price tags and marketing spiel than solutions which simply ‘do the job and do it well’.

Put simply, shop around. Never accept a solution without face-to-face discussions with a technical architect. Come in and see us. Discuss your requirements with our technical team. You can even try to beat us at pool.

Rob Greenwood, Technical Architect.


A brief word on honesty

November 4th, 2009

It seems to be de rigeur in the hosting industry to upsell one’s infrastrucutre, in the process potentially making slightly misleading claims about datacentre and/or network ownership.

Hopefully you wouldn’t expect this kind of behaviour from Melbourne, and rightly so.  All the literature about our network and datacentre specifications is correct, accurate and factual.

Rest assured that:

  • We own and operate our own Manchester Datacentres, totalling approximately 6000 sq.ft. and looking after over 2500 servers.  We don’t put your servers in other people’s datacentres.
  • We own and operate our own service provider network which load-balances connections from multiple providers.  We operate our own 10Gbps fibre ring across manchester, spanning six datacentres.  We don’t use someone else’s network and pass it off as our own.
  • Our support team is based in our Manchester office, adjacent to our two main datacentres.  The same support team answers your calls, replies to your tickets, and works on your servers in the datacentres.

When we’ve gone to so much effort and expense to get these things in place, it does occasionally annoy us when a competitor does the equivalent of sock stuffing, where they’ve not gone to the effort and expense of doing something but claim they do.

Of course you can always find this out for yourself by coming to visit us.  If you can’t make it, why not have a quick look at our “ten reasons that melbourne is a better hosting provider” document, and come to your own conclusion.

Daniel Keighron-Foster, Technical Director.


Dedicated or Virtual?

October 29th, 2009

Now virtualisation has reached maturity, virtual servers are a real alternative to dedicated servers.  However, they’re not ideal for all situations.

From our point-of-view we’re seeing virtual servers as a great alternative to our entry-level dedicated servers, and a great deal faster than older dedicated servers running on deprecated hardware.

That said, they’re not perfect for everything, and there are many applications where only dedicated will suffice.

Here are some of the aspects of virtual servers to consider:

  • CPU is burstable. With a virtual server, you’re allocated a smaller slice of a higher-end CPU than you’d have in a low-end server.  In our case, we’re using Xeon L5420s  and L5520 processors in our virtual server nodes.  These are faster by quite a stretch than the CPUs found in our entry-level servers.  Obviously you don’t have all of this power at your disposal, but  when there are idle cycles, you are free to grab them.  And given that the average CPU usage on our VM nodes is < 10% on average, the end result can be much faster than that 3-year-old chip in that cheap dedicated.
  • I/O is burstable. Slow disk IO is usually the most criticised short-coming of VPS, but it only happens on overcrowded boxes.  In our case, you’re connected to a SAS or SATA SAN (dependant on which storage you pay for) which is running RAID50, connected by a 2Gbps iSCSI connection to the host node.  The End result? Much faster disk I/O with our SAS option, when compared to the disks you find in a low-end dedicated server.
  • Better Resiliency. Compared to a cheap dedicated, you’re getting a slice of a machine which has redundant power supplies, fed from two UPSs, dual uplinks to both frontend and storage networks, and RAID50 storage, with hot-spare disks on our high-availability SAN (with off-site replication too).
  • Better for the environment. Because hardware is better used, less electricity is consumed per virtual server.

Good candidates for virtualisation:

  • Servers which have low CPU and/or memory usage, but with occasional peaks.
  • Servers which can be scaled horizontally, for example web servers behind a load-balancer.
  • Single-service servers, i.e. those running a single website or email.
  • Servers which need to have CPU and/or memory upgraded for short periods, without excessive downtime associated with hardware upgrades.
  • Staging environments needing a fraction of the resources of the live environment they mirror.
  • DR environments which usually need very little resource but need to be scaled upwards quickly.

Bad candidates for virtualisation:

  • Servers needing access to multiple CPUs or masses of memory (i.e. the largest VM we offer as standard has 4GB of RAM and 2 CPU cores).
  • Servers needing huge amounts of storage space.  This is due to the high cost of SAN-based storage.  Once you get above a certain threshold, it becomes cheaper to use a dedicated server fitted with a RAID5 SAS disk array.

Mix-and-match

The great thing about the way we can provision servers means that your virtual and dedicated servers can sit on your own private network (VLAN), and communicate with each other freely.  This means that you have the flexibility to use the right mix of servers based on the individual requirements of each element of your hosting platform.

Our technical team, in conjunction with our account managers, can help to advise on what’s right for any given situation.


Come see us at the Business North West Exhibition

October 27th, 2009

Come and say hello!

We’ll be at the two-day Business North West Exhibition (28th and 29th October) at Manchester Central. The event is free to attend (you can register for tickets here) and is designed for business owners, directors and entrepreneurs running small and medium-sized businesses.

You can see us either on the Manchester Science Park stand (123) or the Networking4Business.com stand (17).  If you’d like us to make a specific appointment to see you, please call our sales team on 0800 975 8771 or 0161 232 0001.


We’re looking for a New Business Sales Exec

October 20th, 2009

Melbourne is on the lookout for a New Business Sales Executive to join our growing team.  The role involves selling our range of managed dedicated servers, managed services and other products and solutions across market sectors.

The position is based from our Hulme, Manchester office.

Click here for details: New Business Sales Executive.


Free setup on rackspace until 30th November

October 19th, 2009

We have a limited amount of quarter, half and full cabinets available due to our new Turing House Datacentre facility opening in July.

Therefore until the end of November, we’re offering free setup on quarter, half and full cabinets in our Reynolds House facility.

Please contact our sales team for a quotation on rack space.


Introducing Melbourne’s Partner Program

October 19th, 2009

We have lots of partners; that’s companies who use our services, add their value, and then resell to their customers.  Some IT companies, others not so IT company-ish.  What they have in common though is that Melbourne highly values their business.  We’ve always gone the extra mile with our partners, but we decided it was about time to formalise this into a Partner Program.

So what’s in it for our partners?

  • Discount on monthly hosting bills based on your spend.
  • Free virtual machine for internal/development/demonstration purposes.
  • Invitations to exclusive partner events.
  • Datacentre tours for your customers, with use of Melbourne’s meeting facilities.
  • Previews of new products and services that are in testing.

The requirements are simple to join the partner program -  you simply need to maintain a contributable monthly spend of at least £500 per month.  We don’t even mind whether you’re re-selling our services to others or simply using a lot of our services in-house to make up the contributable monthly spend.

To find out more, or to join the partner program, simply visit our Partner Program page.  Once you’ve filled in the (very brief) application form, your account manager will be in-touch to check over your account and ensure that you’re receiving all of the benefits.


Optimising the LAMP Stack - MySQL

October 16th, 2009

As with any web application, a great level of thought and planning goes into the optimization of the webapp to ensure fast response times. However, what often gets overlooked is the configuration of the LAMP stack itself. There is plenty of benefit to be gained from optimizing the configuration of Apache, PHP and MySQL.

Although all the different components provide ‘optimised’ configurations, they’re generalised and not written for an applications specific requirements. The aim of the ‘Optimising the LAMP Stack’ series is not to provide an example configuration, but to explain the more important configuration options and what impact those settings have on your web applications - giving a greater in site into how to configure it correctly.

Your MySQL configuration can generally be found in /etc/my.cnf - heres a list of what I see as the most important variables.

  1. query_cache_size:
    • MySQL 4 onwards provides query caching, whereby if a database has to continually run the same queries on the same dataset it will cache the results. Although this level of caching should generally be taken care of by the application, this is a good alternative and saves the server repeating tasks.
  2. thread_cache:
    • Thread creation/destruction can be expensive, and this happens on every new connection. Keep this in line with the amount of connections you expect your  server to receive. If you start to see a spike in threads created, increase it. The goal is to not have threads created during normal operation.
  3. key_buffer_size:
    • The key buffer is used with indexes. The larger you set the buffer, the quicker queries will complete and return a result. I recommend setting this somewhere between a quarter, but no more than half of the systems totally memory. In an ideal situation, this will be sufficient to contain all your table indexes.
  4. table_cache:
    • Each time MySQL opens a table, it’s placed in the cache. The more tables you open, the higher cache you’ll require. Also bare in mind that MySQL is multi-threaded, therefore you could end up opening the same table multiple times in different processes and need to account for this.

The above list is the most important options for optimising a standard MySQL servers workload. Although there are plenty more, a lot can degrade performance when used incorrectly or when an applications requirements differ.

Should you wish to further optimize your configuration, I would suggest looking through the sample ‘optimized’ configuration provided with the MySQL server (Generally located in /usr/share/mysql). MySQLTuner is also pretty useful for configuration reccommendations when your database has been up and running for a few days.

Rob Greenwood, Technical Lead


The five dangers of virtual servers

October 16th, 2009

Virtual servers and cloud computing are changing the way we think about hosting high performance and high-availability applications.  The technology uses virtualisation techniques to abstract the underlying hardware from the end product, i.e. the server itself.

In theory this means:

  • Higher availability as there is no single point of failure (SPF) associated with any particular hardware;
  • Lower costs, as an entire physical node is not dedicated to one server task any more, i.e. multiple virtual servers can run concurrently on one physical node;
  • Lower energy consumption per server, for the same reasons as above;
  • Quicker lead times on server upgrades as “virtual hardware” can be scaled up or down instantly or near-instantly.

Like everything though, the devil is in the detail, and there’s plenty of scope for hosting service providers to get it wrong.

That’s why we’ve put together a white-paper “the five dangers of virtual servers”.  It highlights areas where providers can end up compromising the security or availability of your virtual server, by cutting corners, or badly thinking out the infrastructure when they put it together.

Click here for our exclusive white-paper, “the five dangers of virtual servers”.

And now a little about Melbourne…

Started in 2000, Melbourne now owns and runs three datacentres totalling approximately 6000 sq.ft. of datacentre space and over 2,000 servers. Our customers include The Chartered Institute of Tax, JJB Sports PLC, Travel Counsellors and Freedom Finance PLC.

Obviously web sites and ecommerce applications must be available 24×7x365 year on year so Melbourne provides redundancy on practically every part of the server, power supply and Internet access provision.

The data centres all operate with multiple internet connectivity providers to ensure Internet redundancy. All servers have dual power supplies and with the firm intention to stay in play even if
there’s a prolonged power outage, as Melbourne has a regularly tested UPS and generator set on standby that can power the entire facility for 7 days by itself.

Melbourne’s Virtual Dedicated Server offering, called UltraVM™, was launched in early 2009, providing robust reliability with an affordable price.

If you’ve any questions about this report or want to have a tour of our Manchester data centres please phone us on 0800 915 8771.