17/11/2010

The Situation:

Ozsale is Australia’s first and exclusive private online shopping club which works through offering designer goods at highly discounted rates to people who have signed up as members. There are two tiers of membership, ‘Free’ and ‘Priority Member’. ‘Free’ membership is exactly that, free. There is no membership fee and members are able to purchase online. For a small fee, ‘Priority Members’ receive early access to sales and discounts on the shipping fees.

The success of the business lies in the frequency and volume of sales. The business model revolves around stock moving very quickly, there are new sales every day and stock is only available for 2-3 days, instilling a sense of urgency. Often stock is sold out during the first day of the sale and members know that if they see something they want, they need to buy it quickly. And they do. This also means that members look at the site every day as new stock is released.

The entire business relies on the quality of the infrastructure with the site receiving peak loads at the beginning of sales and then plateauing out for the rest of the day. The infrastructure needs to be robust enough to cope with the spike of peak demands. Not only are there peaks of visitors to the site, but the membership itself is growing at an exponential rate.

The Challenge:

When the business was launched in Australia it was hosted overseas, but it soon became apparent that as the business grew, the performance became worse and worse with server latency issues and Ozsale had clearly outgrown the capability of their initial infrastructure. They were recommended to ICO through a business contact and after discussions on the best approach, entrusted the critical back-end infrastructure to ICO in 2007.

Due to the nature of the business and the sense of urgency that Ozsale has created around their sales, the site receives heavy spikes of activity each morning after the newsletter has been sent out. It was important that the infrastructure had the capacity to handle this unusual pattern of demand.



Initially the solution comprised of 2 high-end physical dedicated servers which suited the business at the time. One was a web server, the other a database server. There was no redundancy for either server. However ICO identified that with their continuing success, Ozsale would need to upgrade their infrastructure to cope with not only the increase in sales but also the expanded online presence. It was important to be proactive and prepare the back-end infrastructure for the rapid growth, enabling the migration to be done without any major impact to the business, rather than as a reactive approach done in an emergency.

“We have always been most impressed with how ICO builds relationships with their customers. Our Account Manager is always very helpful and offers experienced advice for future growth strategies and scalability. We are growing rapidly. We need to know that the company responsible for the back-end of our business is able to support that and we certainly have that in ICO.” Gavin Cliffe, IT Manager, Ozsale

The Solution:

ICO proposed that Ozsale move to a clustered virtualized infrastructure which would not only be highly scalable but would also have complete auto fail-over and redundancy capabilities to ensure against any potential hardware failures which would impact the business. This clustered solution is perfect for a business that could not afford to have any outage.

All of Ozsale’s servers were virtualized and migrated onto their own dedicated cluster of hosts, fully managed by ICO and residing in ICO’s ASIO T4 rated data centre. The virtualized environment is run on VMware and comprises all aspects of VSphere4 including vMotion for live migration. Essentially, should any of the redundant hosts fail, all of the dedicated virtual servers will migrate to the remaining hosts with no loss of data or impact to the business.

“We are absolutely happy with ICO and the new virtualized environment. We have the Level II Management service which is excellent. It was great when our ICO Account Manager contacted us to suggest we start looking at other solutions. They are always proactively working with us and ensuring that we are on the best infrastructure for our business. The ICO Support Team runs on an excellent system which ensures that there is no queuing and waiting for responses. Calls are answered and any problems are worked on immediately to ensure minimum impact for the customer. This makes a massive difference to our business.” Gavin Cliffe, IT Manager, Ozsale

The underlying host infrastructure is completely redundant and has the capability to scale with the business. As with all of ICO solutions, Ozsale’s new infrastructure has a 99.999% network uptime guarantee and 24x7 immediate response Support.

Now expansion is a simple case of deploying additional virtual servers without any impact to the cluster as a whole.

“It’s great being able to call and get through to the right person. All of our software licensing requirements were sorted out by ICO which is great as this is usually out of the scope of ‘hosters’. The main thing for us is that ICO are always working with us to look at how we can increase capacity to meet the expanding business.” Gavin Cliffe, IT Manager, Ozsale
 

Posted by Rachel Holden | with 0 comments
01/09/2010
If you are signed up to our newsletter you will already know the topic of this blog! You can subscribe on every page of the website, in fact, scroll down to the bottom of this page now to subscribe.
 

Planning for DRBC


It’s one of those things that you know you need to implement, but every day emergencies get in the way and putting your, in effect, ‘insurance plan’ into place often takes a back seat, leaving your business open to the serious risk of downtime and data loss. Of course, there are many of you who do put a plan into place, and it may not be as costly or onerous as you originally had expected...

 

The Challenge:
Taurus needed a Disaster Recovery (DR) solution in order to safe guard against any unforseen eventualities and also to adhere with the necessary compliance requirements of the Australian Financial Services regulatory body, ASIC. Previously the company had an offsite back-up routine but soon wanted a more robust solution which would enable the company to continue running with minimal downtime in the event of a disaster.

 
The Solution:
“The beauty of this system is that we know it is being maintained by ICO and it is ready whenever we need it, in the event of an emergency.”
 
The DR Solution comprises of (click to find out!)
 
In the event of a disaster, terrorist attack, power cut, or even restricted access to the office building Taurus is able to continue working until the issue is resolved. This isn’t a theoretical solution; the system is tested annually to assimilate a disaster.
 
As always, thank you for taking the time to read our blog!
Posted by Rachel Holden | with 0 comments
27/07/2010

How to Configure Your Mail Server Correctly!

To view the beginning of this blog, sign up to our newsletter (enter your email address in the top right of our homepage)!

Now, you've followed the previous instructions and have ended up with 6 tabs:

SMTP

Add 127.0.0.1 to the allowed relay list as follows:

  • Click on the 'Access' tab
  • Click 'Relay'
  • Ensure the "Only the list below" radio button is checked
  • Click 'Add' and type 127.0.0.1 (localhost)
  • Click 'OK'

Relay Restrictions

Now test your email again. This time it works! However, it may have been classified as 'junk mail'. Clearly you want to improve the chances of delivery, avoiding your email ending up in the junk mail folder. It's actually a RFC requirement for a mail server to be resolvable on the Internet. RFC 2821 2.3.5 - http://www.ietf.org/rfc/rfc2821.txt

Stop Your Email Going to Junk Mail

To do this you will need another DNS A record for your mail server. The next step is to enter the DNS A record into the SMTP server:

  • Again, going to the Properties of the SMTP server
  • Selecting the 'Delivery' tab
  • Select 'Advanced'
  • Enter the A record you have created (and pointed to your server) in the 'Fully Qualified Domain Name' space
  • Click 'OK'

Advanced Delivery

Once this is done the next step is to email ICO Support with your current DNS A record requesting the creation of a PTR (Pointer or Reverse DNS entry).

The use of a DNS PTR record in mail delivery is to verify that the sending mail server is who it claims to be by resolving the IP address to its A record.

Simple Mail Flow

And that's how you set up your Mail Server and avoid your email losing its way!

Posted by Rachel Holden | with 0 comments
16/02/2010

At ICO we are passionate about our support and customer service, not just from our highly knowledge technical team, but throughout the business, from sales and marketing to operations.

Our prime goal is to always add value for new prospective clients (and our existing clients), from the perspective of offering not only a sound technical solution, but also a cost effective solution.

We recently received an enquiry from a new client who was running two, three year old servers in a Melbourne based data center, under a Co-Location arrangement.

The client, who was based in Sydney, initially asked if ICO could provide a Co-Location arrangement for these servers in our Sydney data center, as the arrangement in Melbourne was becoming very problematic from an ongoing administration and support perspective.

Whilst ICO does not generally offer Co-Location services, during the course of the initial discussion we became aware this client faced some very significant problems in managing their IT environment, namely:-
 

  • The existing servers were quite old and the customer was faced with a looming decision to incur additional high cost capital expenditure to replace them
     
  • They had to carry the cost for 2 servers because the business model was critical and therefore they needed to run a replicated “hot standby” server to cover the potential of a production server failure 
  • They were faced with the challenge of finding a competent organisation who could provide backup support/maintenance for the hardware and software
  • Apart from their own specialist supply chain management application, they used a Plesk control panel to manage multiple web sites and a Mail Enable server for email.  Day to day management and administration of both Plesk and Mail Enable was a major problem for them as they had no one on staff capable of handling this role.  Subsequently, they had suffered periods of downtime in the past due to a lack of local technical knowledge on these products. 

Once the scale of the clients challenges became clear, our account manager was able to recommend an approach that resolved all the issues in a very cost effective manner.  For less than $800/mth, we were able to offer this client (and they accepted) a fully managed hosting package covering:

  • A fully managed server running on our VMWare ESX Infrastructure,
  • live support engineer 24 x 7 support,
  • full dynamic failover capability (dispensing with the client’s need to run 2 dedicated servers),
  • full daily managed backup,
  • full Anti Virus management and
  • complete management of all Operating System software.
  • An ICO fully managed service to look after the administration and management of both Plesk and Mail Enable.
Posted by Global Administrator | with 0 comments
21/01/2009

I know there are a lot of blogs about the twitter phenomenon and I don't intend to turn our blog into a twitter blog. However, as early adopters of technology I thought it was worthwhile to talk about our twitter experience.

It seems that more and more individuals and organisations are using twitter. For example, the new USA President, President Obama has used twitter as part of his campaign and continues to use it to get messages out to thousands of people. And closer to home, Mosman Council, a local sydney council, has become quite active recently and even encourages local residents to submit photos and get involved in the 'Community Conversation'. Reports have even been written about twitter, O'Reilly Media produced 'Twitter and the Micro-Messaging Revolution: Communication, Connections and Immediacy - 140 Characters at a Time'.

Our own experience of twitter has been positive. Our internal emails have significantly reduced and internal communication has increased, especially with our team members who work offsite.

Twitter has been great for improving communication, even if the comments posted are not directly about work, it's good to know what other people are doing. It's also a lot quicker than sending an email and is completely acceptable to post a short, sharp message. No one is obliged to reply.

We use it for feedback on an idea, and will use it to prompt brain storming sessions. It has definitely added a 'buzz' to the team.

Of course there are downsides of applications traditionally known as 'social networking'. People can get immersed in them and spend hours updating their status and photos. Fortunately we haven't had this problem with twitter as we have guidelines for use which encourage updating but also give a daily time limit. Also, when limited to 140 characters there is less likelihood of excessive time being spent writing updates. There's only so much you can say in 140 characters!

We have also incorporated our twitter IDs into our email signatures and they will also appear in the contact details of our new business cards.

Feel free to comment on this blog. What do you think about twitter? Is it a waste of your time or do your colleagues, family and friends appreciate the quick updates?

Posted by Global Administrator | with 0 comments
  Blog postsRSS

 

An Avnet Company

ICO Uses & Recommends

MicrosoftVMwareRed HatIntelSunMail EnableCiscoFortinet

Sales

1300 426 316
enquiry@ico.com.au

Support

1300 727 733
support@ico.com.au