Emulex is driving convergence in the data centre

Our last but one visit in the Silicon Valley on November 18 was with Emulex, and we were greeted by Steve Daheb, CMO & SVP business development and Shaun Walsh, VP Corporate Marketing


The Emulex vision is about the fact they realised that networks were about to converge for the first time:  data networks used to be separated from storage networks. With the event of 10Gb ethernet we are seeing things converge. Emulex is a 30-year company founded in 1978, based in Costa Mesa, Calif. with 960 employees. Emulex works with HP, IBM and all other major companies. Emulex is also a great partner of Blade Network Technologies’ which we have visited previously.

There are 2 product lines at Emulex: 1) host servers 2) back-end storage which is a legacy business for Emulex. Emulex has a legacy business in Fibre channel (a technology used to build SANs in which Cisco and Brocade are leaders).

Recently, Emulex acquired ServerEngine, which with they have worked for 2 years. They closed the deals with HP and IBM with ServerEngine and Emulex is winning quite a few deals in which their chips are equipping OEM 10 Gb Ethernet equipment and 9 out of 10 tier 1 OEMs have picked them for convergence (some exclusively, some not), amongst which HP, Dell, Cisco, Fujistu Siemens and IBM. A big part of the company is already working on what will come next and will be delivered in 2012 and 2013.

the data centre of the future

A lot of enterprises are still at the crossroads with data, email and storage networks all considered 3 distinct networks. With this kind of architecture, customers are compelled to upgrade their data centres and cost is a big issue. A second step is to “move everything into virtual data centres and this is where we are today” Shaun added.

But in the last 24 months, we have seen the rise of converged fabric-based computing as announced by HP (Flexfabric), IBM (virtual fabric) and Cisco. Convergence is happening, for server virtualisation, switching, storage and adapters (LOMs i.e. LAN on Motherboard).  These past 24 months, convergence has been flavour of the month and is showing the way forward.

from servers to blades and lastly … to “cookie sheets

The 10Gb Ethernet is a new inflection point which is a sign for innovation, and such innovation is made flesh with 4 main trends which Steve described here. These 4 trends are defining the future of data centres:

  1. data is growing at an increasing rate and this is driving huge demand in bandwidth,
  2. IT priorities are based around how one can deploy more quickly,
  3. cost drivers are prevalent: lower capex/opex and data centre consolidation,
  4. technology trends such as “cloud computing” which necessitate a convergent environment, mostly for SMEs.

Behind these 4 trends, Steve described 3 business drivers which are impacting the way that data centres are being architected.

  1. SMBs an remote office data centres
  2. enterprise hybrid data centres: a mixture of internal data centres for data privacy and public cloud applications for cost reasons
  3. hyper scale data centres

Data is growing at a staggering compound annual growth rates of approx. 50%. More users mean more transactions and convergence is therefore required. 9m servers will be shipped in 2010, 1m+ will be of the new breed, Shaun said and he added that we are moving from servers to blades and now what Shaun called cookie sheets because they are so thin.

from a networks point of view, what is driving change?

Steve described the 5 reasons for change in the data centre from the network perspective:

  1. the explosion of devices which will eventually lead to the replacement of the laptop, and the explosion of content and transactions,
  2. new technology is being brought to the edge – as the new stuff being brought by Riverbed – and more intelligence is needed in the data centre,
  3. video, web-enabled mobile is boosting traffic so much so that switch port shipments are rocketing up wit happrox. 60% CAGR (see above photo),
  4. security is also a huge driver,
  5. network traffic is not only driven by consumers but also enterprises, with mobile acces to applications on mobiles.

the data centre of the future requires a new networking model

Emulex sees his role as that of a leader in that move into convergence, partnering with BNT, VMWare and all their OEMs.

the management mind meld book available for download

If you want to know more about the Emulex vision, we recommend you download their online e-book “the management mind meld”

you can also follow Emulex on Twitter at @emulex

Converged networking results in an overlap of network and storage administrators’ responsibilities. This guide explains networking and storage basics to help each administrator better understand the changes resulting from converged networking and how it will impact their role in the data center.

Improving management’s vision of the realities of business : Processes or Networks?

improving Management's vision of business realitiesIn response to our initial post, C.R. asked for more details on how IT distorts management’s vision of reality. Given the work and resources that have gone into enterprise applications over the last two decades, he asks, why doesn’t IT feed back a true image of how we work? If the daily statistics and scorecards provided by today’s business applications don’t seem to “ring true”, what can an organization do to improve the quality of the data to better reflect individual customers, projects, and opportunities? To understand this challenge, and to explore potential scenarios to improve enterprise applications, we need to first consider how business applications today represent and aggregate data, then explore the fundamental weaknesses in this logic, and finally investigate other avenues of putting the pieces together.

Enterprise applications, including those dedicated to enterprise resource planning, supply chain management and even project management, and are largely based on the principles of business process improvement. A process can best be understood as a set of activities and tasks that managers group together to meet internal or external client demands. Processes thus have four defining characteristics: their origin can be detected in an organization’s recognition of a client need for a product, service or information, they consume resources, and they have a set cost and produce a benefice that ideally meets the clients’ needs.

Improving business processes requires tracking inefficiencies in how an organization captures market demand, supply or measurement. How are an organizational activities and tasks executed to capture client demands? How can an organization improve its capacity to deliver products and services? How does the organization capture and evaluate costs and benefits? Enterprise applications provide responses to these fundamental questions in allowing management to map the current state of existing processes, implement best practices around ideal designs of how these processes should work, and measure organizational progress to these goals.

The resulting ‘process-centric” view of the organization is often a weak approximation of the reality of business practice, and as such limits the usefulness of enterprise applications. One reason for this is that business process improvement was originally designed for organizations producing standardized physical goods, modeling processes around personalized services and/or information delivery has proven a much more difficult assignment. Implementing business processes assumes that management (or their business partners) understand where an industry or a market is headed, an assumption that is sorely tested today in industries ranging from information services to banking to telecommunications.

Moreover, the notion of “best practices” which assumes that there is “one best way” to build product or deliver services has been widely disputed in organizations that have traditionally put a premium on company culture, customization, and understanding individual customer needs. Finally, improving business processes assumes that the employees and managers involved in the targeted processes agree with organizational objectives, and are willing and able to follow organizational guidelines. In many situations, even this assumption is contested by those that prefer focusing on the realities of their organization and their market (i.e. getting the job done) than complying to unrealistic or unworkable company procedures. For all of these reasons, “process-centric” software most often offers management a highly distorted view of the realities of the workplace.

If we abandon the notion of processes, how can business applications be modeled to provide a more realistic view of work? A growing number of management specialists are underlining the importance of networks in understanding how employees and managers actually get work done. In business, networks are interconnected systems of people that share common interests, beliefs, and goals. Social and business networks are used by workers and managers alike to solve problems, to identify opportunities, to build trust and passion, and make sense of their jobs, organizations and careers. These networks cross organizational boundaries; they are composed of centers of influence (the “hubs”) and are held together by the intensity of personal relationships (the “links”).

If most managers understand the importance of their personal networks, improving networks requires understanding how these networks function and how they can be extended and strengthened. Social researchers have identified a number of defining characteristics of networks that potentially can help mangers improve their effectiveness: “power laws“, “degrees of separation”, “discontinuous change“, “fat tails“, etc. These networks are held together by common beliefs, passions, trust or know-how: in other words non-structured information that largely escapes “process-centric” view of enterprise. The power of networks has fueled the popularity of a new breed of the social media applications: LinkedIn, Plaxo, and Facebook are among the better known examples. Will this new generation of software applications help management improve their business?

One value proposition of such network-centric applications is to shift the focus of attention away from an ideal set of activities and tasks (what we “should” be doing) to how employees and managers actually get things done (what they “are” doing). A second insight of a network based view of business is the understanding that networks are self-structuring; people seek to work with those their share their goals, passions and beliefs. A third point of interest is understanding the importance of “non-structured” data, just because we can’t get passions, trust and know-how into a spreadsheet doesn’t mean they don’t largely influence the way we work. Finally, this line of inquiry reminds us that business is essentially a human activity; technology’s role in business is limited to understanding, uncovering and eventually enhancing the human interaction that defines the nature of “work”. For management and their business partners, these propositions constitute powerful levers to improve the value of information technology.

None-the-less, there certainly is a considerable amount of work to be done before applications like LinkedIn or Facebook help managers improve their jobs, organizations and careers. The metaphors of “avatars”, “friends” and “connections” don’t easily translate into the realities of either commerce or industry. The personal information, photos, and videos available in social media today are certainly interesting, but often of little use in driving a business forward. These applications today are largely hard coded; it is difficult or often impossible for end-users to enrich either the information or the software. These applications capture data on how end-users wish to see themselves, rather than how they actually practice business (or anything else…). Most importantly, these virtual networks don’t elucidate today the passions, beliefs, and goals that define how individual professional networks operate. If network-centric applications provide a glance of how the future of enterprise applications can improve business practice, these images need to be brought into focus.